Make a start on the JIT IR and the trace builder.

The JIT IR is designed to be small. There are two kinds of instruction:
 - short instructions, with inlined operands.
 - long ones (unimplemented as of yet)

Ideally, function, block and instruction IDs (which are indices read
from the on-disk AOT IR) would be either converted to references as we
decode, but this would require changes to the decoder. We may have to
ditch deku:

sharksforarms/deku#383

Co-authored-by: Lukas Diekmann <[email protected]>

ykrt/Cargo.toml

@@ -19,6 +19,7 @@ ykaddr = { path = "../ykaddr" }
 yksmp = { path = "../yksmp" }
 strum = { version = "0.25", features = ["derive"] }
 yktracec = { path = "../yktracec" }
+strum_macros = "0.25.3"
 
 [dependencies.llvm-sys]
 # note: using a git version to get llvm linkage features in llvm-sys (not in a

ykrt/src/compile/jitc_yk/aot_ir.rs

@@ -5,7 +5,15 @@
 
 use byteorder::{NativeEndian, ReadBytesExt};
 use deku::prelude::*;
-use std::{cell::RefCell, error::Error, ffi::CStr, fs, io::Cursor, path::PathBuf};
+use std::{
+    cell::RefCell,
+    error::Error,
+    ffi::CStr,
+    fs,
+    io::Cursor,
+    ops::{Deref, DerefMut},
+    path::PathBuf,
+};
 
 /// A magic number that all bytecode payloads begin with.
 const MAGIC: u32 = 0xedd5f00d;
@@ -45,7 +53,7 @@ pub(crate) trait IRDisplay {
 
 /// An instruction opcode.
 #[deku_derive(DekuRead)]
-#[derive(Debug, Eq, PartialEq)]
+#[derive(Clone, Copy, Debug, Eq, PartialEq)]
 #[deku(type = "u8")]
 pub(crate) enum Opcode {
     Nop = 0,
@@ -82,14 +90,53 @@ impl IRDisplay for ConstantOperand {
 }
 
 #[deku_derive(DekuRead)]
-#[derive(Debug)]
-pub(crate) struct LocalVariableOperand {
+#[derive(Debug, Hash, Eq, PartialEq)]
+pub(crate) struct InstructionID {
     #[deku(skip)] // computed after deserialisation.
     func_idx: usize,
     bb_idx: usize,
     inst_idx: usize,
 }
 
+impl InstructionID {
+    pub(crate) fn new(func_idx: usize, bb_idx: usize, inst_idx: usize) -> Self {
+        Self {
+            func_idx,
+            bb_idx,
+            inst_idx,
+        }
+    }
+}
+
+#[derive(Debug)]
+pub(crate) struct BlockID {
+    pub(crate) func_idx: usize,
+    pub(crate) bb_idx: usize,
+}
+
+impl BlockID {
+    pub(crate) fn new(func_idx: usize, bb_idx: usize) -> Self {
+        Self { func_idx, bb_idx }
+    }
+}
+
+#[deku_derive(DekuRead)]
+#[derive(Debug, Hash, Eq, PartialEq)]
+pub(crate) struct LocalVariableOperand(InstructionID);
+
+impl Deref for LocalVariableOperand {
+    type Target = InstructionID;
+    fn deref(&self) -> &Self::Target {
+        &self.0
+    }
+}
+
+impl DerefMut for LocalVariableOperand {
+    fn deref_mut(&mut self) -> &mut Self::Target {
+        &mut self.0
+    }
+}
+
 impl IRDisplay for LocalVariableOperand {
     fn to_str(&self, _m: &Module) -> String {
         format!("${}_{}", self.bb_idx, self.inst_idx,)
@@ -161,6 +208,30 @@ pub(crate) enum Operand {
     Unimplemented(#[deku(until = "|v: &u8| *v == 0", map = "deserialise_string")] String),
 }
 
+impl Operand {
+    /// If the operand is a `LocalVariable`, then return the instruction that defines it, or else
+    /// None.
+    ///
+    /// OPT: This is expensive.
+    pub(crate) fn to_instr<'a>(&self, aotmod: &'a Module) -> Option<&'a Instruction> {
+        match self {
+            Self::LocalVariable(lvo) => {
+                Some(&aotmod.funcs[lvo.func_idx].blocks[lvo.bb_idx].instrs[lvo.inst_idx])
+            }
+            _ => None,
+        }
+    }
+
+    /// Return the `InstructionID` of a local variable operand. Must not be called on other
+    /// operands.
+    pub(crate) fn to_instr_id(&self) -> InstructionID {
+        match self {
+            Self::LocalVariable(lvo) => InstructionID::new(lvo.func_idx, lvo.bb_idx, lvo.inst_idx),
+            _ => unreachable!(),
+        }
+    }
+}
+
 impl IRDisplay for Operand {
     fn to_str(&self, m: &Module) -> String {
         match self {
@@ -190,6 +261,43 @@ pub(crate) struct Instruction {
     name: RefCell<Option<String>>,
 }
 
+impl PartialEq for Instruction {
+    fn eq(&self, other: &Self) -> bool {
+        std::ptr::eq(self, other)
+    }
+}
+
+impl Instruction {
+    pub(crate) fn get_operand(&self, index: usize) -> &Operand {
+        self.operands.get(index).unwrap()
+    }
+
+    pub(crate) fn opcode(&self) -> Opcode {
+        self.opcode
+    }
+
+    pub(crate) fn is_store(&self) -> bool {
+        self.opcode == Opcode::Store
+    }
+
+    pub(crate) fn is_gep(&self) -> bool {
+        self.opcode == Opcode::GetElementPtr
+    }
+
+    pub(crate) fn is_control_point(&self, aot_mod: &Module) -> bool {
+        if self.opcode == Opcode::Call {
+            let op = &self.operands[0];
+            match op {
+                Operand::Function(fop) => {
+                    return aot_mod.funcs[fop.func_idx].name == "__ykrt_control_point";
+                }
+                _ => todo!(),
+            }
+        }
+        false
+    }
+}
+
 impl IRDisplay for Instruction {
     fn to_str(&self, m: &Module) -> String {
         if self.opcode == Opcode::Unimplemented {
@@ -249,8 +357,9 @@ impl IRDisplay for Instruction {
 pub(crate) struct Block {
     #[deku(temp)]
     num_instrs: usize,
+    // FIXME: unpub
     #[deku(count = "num_instrs")]
-    instrs: Vec<Instruction>,
+    pub instrs: Vec<Instruction>,
 }
 
 impl IRDisplay for Block {
@@ -546,7 +655,21 @@ impl Module {
         *self.var_names_computed.borrow_mut() = true;
     }
 
-    /// Fill in the function index of local variable operands of instructions.o
+    pub(crate) fn func_index(&self, find_func: &str) -> Option<usize> {
+        // OPT: create a cache in the Module.
+        for (f_idx, f) in self.funcs.iter().enumerate() {
+            if f.name == find_func {
+                return Some(f_idx);
+            }
+        }
+        None
+    }
+
+    pub(crate) fn block(&self, bid: &BlockID) -> Option<&Block> {
+        self.funcs.get(bid.func_idx)?.block(bid.bb_idx)
+    }
+
+    /// Fill in the function index of local variable operands of instructions.
     ///
     /// FIXME: It may be possible to do this as we deserialise, instead of after the fact:
     /// https://github.com/sharksforarms/deku/issues/363
@@ -571,16 +694,12 @@ impl Module {
         &self.types[instr.type_index]
     }
 
+    // FIXME: rename this to `is_def()`, which we've decided is a beter name.
+    // FIXME: also move this to the `Instruction` type.
     fn instr_generates_value(&self, i: &Instruction) -> bool {
         self.instr_type(i) != &Type::Void
     }
 
-    /// Retrieve the named function from the AOT module.
-    pub(crate) fn func_by_name(&self, name: &str) -> Option<&Function> {
-        // OPT: Cache function indices somewhere for faster lookup.
-        self.funcs.iter().find(|f| f.name == name)
-    }
-
     pub(crate) fn to_str(&self) -> String {
         let mut ret = String::new();
         ret.push_str(&format!("# IR format version: {}\n", self.version));