gh-119689: generate stack effect metadata for pseudo instructions (#119691)

5 files changed, 47 insertions, 20 deletions

diff --git a/Tools/cases_generator/analyzer.py b/Tools/cases_generator/analyzer.py
index fdb6354..e44bebd 100644
--- a/Tools/cases_generator/analyzer.py
+++ b/Tools/cases_generator/analyzer.py
@@ -235,6 +235,7 @@ class Instruction:
 @dataclass
 class PseudoInstruction:
     name: str
+    stack: StackEffect
     targets: list[Instruction]
     flags: list[str]
     opcode: int = -1
@@ -295,7 +296,7 @@ def convert_stack_item(item: parser.StackEffect, replace_op_arg_1: str | None) -
         item.name, item.type, cond, (item.size or "1")
     )
 
-def analyze_stack(op: parser.InstDef, replace_op_arg_1: str | None = None) -> StackEffect:
+def analyze_stack(op: parser.InstDef | parser.Pseudo, replace_op_arg_1: str | None = None) -> StackEffect:
     inputs: list[StackItem] = [
         convert_stack_item(i, replace_op_arg_1) for i in op.inputs if isinstance(i, parser.StackEffect)
     ]
@@ -706,6 +707,7 @@ def add_pseudo(
 ) -> None:
     pseudos[pseudo.name] = PseudoInstruction(
         pseudo.name,
+        analyze_stack(pseudo),
         [instructions[target] for target in pseudo.targets],
         pseudo.flags,
     )
diff --git a/Tools/cases_generator/interpreter_definition.md b/Tools/cases_generator/interpreter_definition.md
index 889f58f..ba09931 100644
--- a/Tools/cases_generator/interpreter_definition.md
+++ b/Tools/cases_generator/interpreter_definition.md
@@ -124,7 +124,13 @@ and a piece of C code describing its semantics::
     "family" "(" NAME ")" = "{" NAME ("," NAME)+ [","] "}" ";"
 
   pseudo:
-    "pseudo" "(" NAME ")" = "{" NAME ("," NAME)+ [","] "}" ";"
+    "pseudo" "(" NAME "," stack_effect ["," "(" flags ")"]")" = "{" NAME ("," NAME)+ [","] "}" ";"
+
+  flags:
+    flag ("|" flag)*
+
+  flag:
+    HAS_ARG | HAS_DEOPT | etc..
 ```
 
 The following definitions may occur:
diff --git a/Tools/cases_generator/opcode_metadata_generator.py b/Tools/cases_generator/opcode_metadata_generator.py
index 04fecb2..2632eb8 100644
--- a/Tools/cases_generator/opcode_metadata_generator.py
+++ b/Tools/cases_generator/opcode_metadata_generator.py
@@ -10,6 +10,7 @@ import sys
 from analyzer import (
     Analysis,
     Instruction,
+    PseudoInstruction,
     analyze_files,
     Skip,
     Uop,
@@ -94,12 +95,18 @@ def emit_stack_effect_function(
 def generate_stack_effect_functions(analysis: Analysis, out: CWriter) -> None:
     popped_data: list[tuple[str, str]] = []
     pushed_data: list[tuple[str, str]] = []
-    for inst in analysis.instructions.values():
+    def add(inst: Instruction | PseudoInstruction) -> None:
         stack = get_stack_effect(inst)
         popped = (-stack.base_offset).to_c()
         pushed = (stack.top_offset - stack.base_offset).to_c()
         popped_data.append((inst.name, popped))
         pushed_data.append((inst.name, pushed))
+
+    for inst in analysis.instructions.values():
+        add(inst)
+    for pseudo in analysis.pseudos.values():
+        add(pseudo)
+
     emit_stack_effect_function(out, "popped", sorted(popped_data))
     emit_stack_effect_function(out, "pushed", sorted(pushed_data))
 
diff --git a/Tools/cases_generator/parsing.py b/Tools/cases_generator/parsing.py
index 0d54820..cc897ff 100644
--- a/Tools/cases_generator/parsing.py
+++ b/Tools/cases_generator/parsing.py
@@ -138,6 +138,8 @@ class Family(Node):
 @dataclass
 class Pseudo(Node):
     name: str
+    inputs: list[InputEffect]
+    outputs: list[OutputEffect]
     flags: list[str]  # instr flags to set on the pseudo instruction
     targets: list[str]  # opcodes this can be replaced by
 
@@ -409,16 +411,18 @@ class Parser(PLexer):
             if self.expect(lx.LPAREN):
                 if tkn := self.expect(lx.IDENTIFIER):
                     if self.expect(lx.COMMA):
-                        flags = self.flags()
-                    else:
-                        flags = []
-                    if self.expect(lx.RPAREN):
-                        if self.expect(lx.EQUALS):
-                            if not self.expect(lx.LBRACE):
-                                raise self.make_syntax_error("Expected {")
-                            if members := self.members():
-                                if self.expect(lx.RBRACE) and self.expect(lx.SEMI):
-                                    return Pseudo(tkn.text, flags, members)
+                        inp, outp = self.io_effect()
+                        if self.expect(lx.COMMA):
+                            flags = self.flags()
+                        else:
+                            flags = []
+                        if self.expect(lx.RPAREN):
+                            if self.expect(lx.EQUALS):
+                                if not self.expect(lx.LBRACE):
+                                    raise self.make_syntax_error("Expected {")
+                                if members := self.members():
+                                    if self.expect(lx.RBRACE) and self.expect(lx.SEMI):
+                                        return Pseudo(tkn.text, inp, outp, flags, members)
         return None
 
     def members(self) -> list[str] | None:
diff --git a/Tools/cases_generator/stack.py b/Tools/cases_generator/stack.py
index 5aecac3..7f07a68 100644
--- a/Tools/cases_generator/stack.py
+++ b/Tools/cases_generator/stack.py
@@ -1,7 +1,8 @@
 import re
-from analyzer import StackItem, Instruction, Uop
+from analyzer import StackItem, StackEffect, Instruction, Uop, PseudoInstruction
 from dataclasses import dataclass
 from cwriter import CWriter
+from typing import Iterator
 
 UNUSED = {"unused"}
 
@@ -208,13 +209,20 @@ class Stack:
         return f"/* Variables: {[v.name for v in self.variables]}. Base offset: {self.base_offset.to_c()}. Top offset: {self.top_offset.to_c()} */"
 
 
-def get_stack_effect(inst: Instruction) -> Stack:
+def get_stack_effect(inst: Instruction | PseudoInstruction) -> Stack:
     stack = Stack()
-    for uop in inst.parts:
-        if not isinstance(uop, Uop):
-            continue
-        for var in reversed(uop.stack.inputs):
+    def stacks(inst : Instruction | PseudoInstruction) -> Iterator[StackEffect]:
+        if isinstance(inst, Instruction):
+            for uop in inst.parts:
+                if isinstance(uop, Uop):
+                    yield uop.stack
+        else:
+            assert isinstance(inst, PseudoInstruction)
+            yield inst.stack
+
+    for s in stacks(inst):
+        for var in reversed(s.inputs):
             stack.pop(var)
-        for i, var in enumerate(uop.stack.outputs):
+        for var in s.outputs:
             stack.push(var)
     return stack