Add integer binary instructions to Winch (bytecodealliance#6538)

* Add integer binary instructions to Winch

* Use handle_invalid_operand_combination and load_constant

fuzz/fuzz_targets/differential.rs

@@ -328,6 +328,22 @@ fn winch_supports_module(module: &[u8]) -> bool {
                         | I64GeU { .. }
                         | I32Eqz { .. }
                         | I64Eqz { .. }
+                        | I32And { .. }
+                        | I64And { .. }
+                        | I32Or { .. }
+                        | I64Or { .. }
+                        | I32Xor { .. }
+                        | I64Xor { .. }
+                        | I32Shl { .. }
+                        | I64Shl { .. }
+                        | I32ShrS { .. }
+                        | I64ShrS { .. }
+                        | I32ShrU { .. }
+                        | I64ShrU { .. }
+                        | I32Rotl { .. }
+                        | I64Rotl { .. }
+                        | I32Rotr { .. }
+                        | I64Rotr { .. }
                         | LocalGet { .. }
                         | LocalSet { .. }
                         | Call { .. }

winch/codegen/src/isa/aarch64/masm.rs

@@ -8,7 +8,10 @@ use crate::{
     abi::{self, local::LocalSlot},
     codegen::CodeGenContext,
     isa::reg::Reg,
-    masm::{CalleeKind, CmpKind, DivKind, MacroAssembler as Masm, OperandSize, RegImm, RemKind},
+    masm::{
+        CalleeKind, CmpKind, DivKind, MacroAssembler as Masm, OperandSize, RegImm, RemKind,
+        ShiftKind,
+    },
 };
 use cranelift_codegen::{settings, Final, MachBufferFinalized};
 
@@ -179,6 +182,22 @@ impl Masm for MacroAssembler {
         self.asm.mul(rhs.into(), lhs.into(), dst.into(), size);
     }
 
+    fn and(&mut self, _dst: RegImm, _lhs: RegImm, _rhs: RegImm, _size: OperandSize) {
+        todo!()
+    }
+
+    fn or(&mut self, _dst: RegImm, _lhs: RegImm, _rhs: RegImm, _size: OperandSize) {
+        todo!()
+    }
+
+    fn xor(&mut self, _dst: RegImm, _lhs: RegImm, _rhs: RegImm, _size: OperandSize) {
+        todo!()
+    }
+
+    fn shift(&mut self, _context: &mut CodeGenContext, _kind: ShiftKind, _size: OperandSize) {
+        todo!()
+    }
+
     fn div(&mut self, _context: &mut CodeGenContext, _kind: DivKind, _size: OperandSize) {
         todo!()
     }

winch/codegen/src/isa/x64/asm.rs

@@ -2,7 +2,7 @@
 
 use crate::{
     isa::reg::Reg,
-    masm::{CalleeKind, CmpKind, DivKind, OperandSize, RemKind},
+    masm::{CalleeKind, CmpKind, DivKind, OperandSize, RemKind, ShiftKind},
 };
 use cranelift_codegen::{
     entity::EntityRef,
@@ -11,7 +11,8 @@ use cranelift_codegen::{
     isa::x64::{
         args::{
             self, AluRmiROpcode, Amode, CmpOpcode, DivSignedness, ExtMode, FromWritableReg, Gpr,
-            GprMem, GprMemImm, RegMem, RegMemImm, SyntheticAmode, WritableGpr, CC,
+            GprMem, GprMemImm, Imm8Gpr, Imm8Reg, RegMem, RegMemImm,
+            ShiftKind as CraneliftShiftKind, SyntheticAmode, WritableGpr, CC,
         },
         settings as x64_settings, CallInfo, EmitInfo, EmitState, Inst,
     },
@@ -59,6 +60,12 @@ impl From<Reg> for GprMemImm {
     }
 }
 
+impl From<Reg> for Imm8Gpr {
+    fn from(value: Reg) -> Self {
+        Imm8Gpr::new(Imm8Reg::Reg { reg: value.into() }).expect("valid Imm8Gpr")
+    }
+}
+
 impl From<OperandSize> for args::OperandSize {
     fn from(size: OperandSize) -> Self {
         match size {
@@ -94,6 +101,18 @@ impl From<CmpKind> for CC {
     }
 }
 
+impl From<ShiftKind> for CraneliftShiftKind {
+    fn from(value: ShiftKind) -> Self {
+        match value {
+            ShiftKind::Shl => CraneliftShiftKind::ShiftLeft,
+            ShiftKind::ShrS => CraneliftShiftKind::ShiftRightArithmetic,
+            ShiftKind::ShrU => CraneliftShiftKind::ShiftRightLogical,
+            ShiftKind::Rotl => CraneliftShiftKind::RotateLeft,
+            ShiftKind::Rotr => CraneliftShiftKind::RotateRight,
+        }
+    }
+}
+
 /// Low level assembler implementation for x64.
 pub(crate) struct Assembler {
     /// The machine instruction buffer.
@@ -292,6 +311,148 @@ impl Assembler {
         }
     }
 
+    /// Logical and instruction variants.
+    pub fn and(&mut self, src: Operand, dst: Operand, size: OperandSize) {
+        match &(src, dst) {
+            (Operand::Imm(imm), Operand::Reg(dst)) => {
+                if let Ok(val) = i32::try_from(*imm) {
+                    self.and_ir(val, *dst, size);
+                } else {
+                    let scratch = regs::scratch();
+                    self.load_constant(imm, scratch, size);
+                    self.and_rr(scratch, *dst, size);
+                }
+            }
+            (Operand::Reg(src), Operand::Reg(dst)) => self.and_rr(*src, *dst, size),
+            _ => Self::handle_invalid_operand_combination(src, dst),
+        }
+    }
+
+    fn and_rr(&mut self, src: Reg, dst: Reg, size: OperandSize) {
+        self.emit(Inst::AluRmiR {
+            size: size.into(),
+            op: AluRmiROpcode::And,
+            src1: dst.into(),
+            src2: src.into(),
+            dst: dst.into(),
+        });
+    }
+
+    fn and_ir(&mut self, imm: i32, dst: Reg, size: OperandSize) {
+        let imm = RegMemImm::imm(imm as u32);
+
+        self.emit(Inst::AluRmiR {
+            size: size.into(),
+            op: AluRmiROpcode::And,
+            src1: dst.into(),
+            src2: GprMemImm::new(imm).expect("valid immediate"),
+            dst: dst.into(),
+        });
+    }
+
+    /// Logical or instruction variants.
+    pub fn or(&mut self, src: Operand, dst: Operand, size: OperandSize) {
+        match &(src, dst) {
+            (Operand::Imm(imm), Operand::Reg(dst)) => {
+                if let Ok(val) = i32::try_from(*imm) {
+                    self.or_ir(val, *dst, size);
+                } else {
+                    let scratch = regs::scratch();
+                    self.load_constant(imm, scratch, size);
+                    self.or_rr(scratch, *dst, size);
+                }
+            }
+            (Operand::Reg(src), Operand::Reg(dst)) => self.or_rr(*src, *dst, size),
+            _ => Self::handle_invalid_operand_combination(src, dst),
+        }
+    }
+
+    fn or_rr(&mut self, src: Reg, dst: Reg, size: OperandSize) {
+        self.emit(Inst::AluRmiR {
+            size: size.into(),
+            op: AluRmiROpcode::Or,
+            src1: dst.into(),
+            src2: src.into(),
+            dst: dst.into(),
+        });
+    }
+
+    fn or_ir(&mut self, imm: i32, dst: Reg, size: OperandSize) {
+        let imm = RegMemImm::imm(imm as u32);
+
+        self.emit(Inst::AluRmiR {
+            size: size.into(),
+            op: AluRmiROpcode::Or,
+            src1: dst.into(),
+            src2: GprMemImm::new(imm).expect("valid immediate"),
+            dst: dst.into(),
+        });
+    }
+
+    /// Logical exclusive or instruction variants.
+    pub fn xor(&mut self, src: Operand, dst: Operand, size: OperandSize) {
+        match &(src, dst) {
+            (Operand::Imm(imm), Operand::Reg(dst)) => {
+                if let Ok(val) = i32::try_from(*imm) {
+                    self.xor_ir(val, *dst, size);
+                } else {
+                    let scratch = regs::scratch();
+                    self.load_constant(imm, scratch, size);
+                    self.xor_rr(scratch, *dst, size);
+                }
+            }
+            (Operand::Reg(src), Operand::Reg(dst)) => self.xor_rr(*src, *dst, size),
+            _ => Self::handle_invalid_operand_combination(src, dst),
+        }
+    }
+
+    /// Logical exclusive or with registers.
+    pub fn xor_rr(&mut self, src: Reg, dst: Reg, size: OperandSize) {
+        self.emit(Inst::AluRmiR {
+            size: size.into(),
+            op: AluRmiROpcode::Xor,
+            src1: dst.into(),
+            src2: src.into(),
+            dst: dst.into(),
+        });
+    }
+
+    fn xor_ir(&mut self, imm: i32, dst: Reg, size: OperandSize) {
+        let imm = RegMemImm::imm(imm as u32);
+
+        self.emit(Inst::AluRmiR {
+            size: size.into(),
+            op: AluRmiROpcode::Xor,
+            src1: dst.into(),
+            src2: GprMemImm::new(imm).expect("valid immediate"),
+            dst: dst.into(),
+        });
+    }
+
+    /// Shift with register and register.
+    pub fn shift_rr(&mut self, src: Reg, dst: Reg, kind: ShiftKind, size: OperandSize) {
+        self.emit(Inst::ShiftR {
+            size: size.into(),
+            kind: kind.into(),
+            src: dst.into(),
+            num_bits: src.into(),
+            dst: dst.into(),
+        });
+    }
+
+    /// Shift with immediate and register.
+    pub fn shift_ir(&mut self, imm: u8, dst: Reg, kind: ShiftKind, size: OperandSize) {
+        let imm = imm.into();
+
+        self.emit(Inst::ShiftR {
+            size: size.into(),
+            kind: kind.into(),
+            src: dst.into(),
+            num_bits: Imm8Gpr::new(imm).expect("valid immediate"),
+            dst: dst.into(),
+        });
+    }
+
     /// Signed/unsigned division.
     ///
     /// Emits a sequence of instructions to ensure the correctness of
@@ -469,17 +630,6 @@ impl Assembler {
         });
     }
 
-    /// Logical exclusive or with registers.
-    pub fn xor_rr(&mut self, src: Reg, dst: Reg, size: OperandSize) {
-        self.emit(Inst::AluRmiR {
-            size: size.into(),
-            op: AluRmiROpcode::Xor,
-            src1: dst.into(),
-            src2: src.into(),
-            dst: dst.into(),
-        });
-    }
-
     /// Compare two operands and set status register flags.
     pub fn cmp(&mut self, src: Operand, dst: Operand, size: OperandSize) {
         match &(src, dst) {

winch/codegen/src/isa/x64/masm.rs

@@ -4,7 +4,9 @@ use super::{
     asm::{Assembler, Operand},
     regs::{self, rbp, rsp},
 };
-use crate::masm::{CmpKind, DivKind, MacroAssembler as Masm, OperandSize, RegImm, RemKind};
+use crate::masm::{
+    CmpKind, DivKind, MacroAssembler as Masm, OperandSize, RegImm, RemKind, ShiftKind,
+};
 use crate::{
     abi::{self, align_to, calculate_frame_adjustment, LocalSlot},
     codegen::CodeGenContext,
@@ -193,6 +195,80 @@ impl Masm for MacroAssembler {
         self.asm.mul(src, dst, size);
     }
 
+    fn and(&mut self, dst: RegImm, lhs: RegImm, rhs: RegImm, size: OperandSize) {
+        let (src, dst): (Operand, Operand) = if dst == lhs {
+            (rhs.into(), dst.into())
+        } else {
+            panic!(
+                "the destination and first source argument must be the same, dst={:?}, lhs={:?}",
+                dst, lhs
+            );
+        };
+
+        self.asm.and(src, dst, size);
+    }
+
+    fn or(&mut self, dst: RegImm, lhs: RegImm, rhs: RegImm, size: OperandSize) {
+        let (src, dst): (Operand, Operand) = if dst == lhs {
+            (rhs.into(), dst.into())
+        } else {
+            panic!(
+                "the destination and first source argument must be the same, dst={:?}, lhs={:?}",
+                dst, lhs
+            );
+        };
+
+        self.asm.or(src, dst, size);
+    }
+
+    fn xor(&mut self, dst: RegImm, lhs: RegImm, rhs: RegImm, size: OperandSize) {
+        let (src, dst): (Operand, Operand) = if dst == lhs {
+            (rhs.into(), dst.into())
+        } else {
+            panic!(
+                "the destination and first source argument must be the same, dst={:?}, lhs={:?}",
+                dst, lhs
+            );
+        };
+
+        self.asm.xor(src, dst, size);
+    }
+
+    fn shift(&mut self, context: &mut CodeGenContext, kind: ShiftKind, size: OperandSize) {
+        let top = context.stack.peek().expect("value at stack top");
+
+        if size == OperandSize::S32 && top.is_i32_const() {
+            let val = context
+                .stack
+                .pop_i32_const()
+                .expect("i32 const value at stack top");
+            let reg = context.pop_to_reg(self, None, size);
+
+            self.asm.shift_ir(val as u8, reg, kind, size);
+
+            context.stack.push(Val::reg(reg));
+        } else if size == OperandSize::S64 && top.is_i64_const() {
+            let val = context
+                .stack
+                .pop_i64_const()
+                .expect("i64 const value at stack top");
+            let reg = context.pop_to_reg(self, None, size);
+
+            self.asm.shift_ir(val as u8, reg, kind, size);
+
+            context.stack.push(Val::reg(reg));
+        } else {
+            // Number of bits to shift must be in the CL register.
+            let src = context.pop_to_reg(self, Some(regs::rcx()), size);
+            let dst = context.pop_to_reg(self, None, size);
+
+            self.asm.shift_rr(src.into(), dst.into(), kind, size);
+
+            context.regalloc.free_gpr(src);
+            context.stack.push(Val::reg(dst));
+        }
+    }
+
     fn div(&mut self, context: &mut CodeGenContext, kind: DivKind, size: OperandSize) {
         // Allocate rdx:rax.
         let rdx = context.gpr(regs::rdx(), self);