[clang] [CIR] Add folders for bit manipulation operations (PR #150235)
Sirui Mu via cfe-commits
cfe-commits at lists.llvm.org
Fri Jul 25 06:02:14 PDT 2025
https://github.com/Lancern updated https://github.com/llvm/llvm-project/pull/150235
>From 169c89a95a5714584a91473c76bf1b9859311bf1 Mon Sep 17 00:00:00 2001
From: Sirui Mu <msrlancern at gmail.com>
Date: Wed, 23 Jul 2025 23:04:09 +0800
Subject: [PATCH] [CIR] Add folders for bit manipulation operations
---
clang/include/clang/CIR/Dialect/IR/CIROps.td | 4 +
clang/include/clang/CIR/MissingFeatures.h | 1 +
clang/lib/CIR/Dialect/IR/CIRDialect.cpp | 121 +++++++++++++++
.../Dialect/Transforms/CIRCanonicalize.cpp | 3 +-
clang/test/CIR/Transforms/bit.cir | 141 ++++++++++++++++++
5 files changed, 269 insertions(+), 1 deletion(-)
create mode 100644 clang/test/CIR/Transforms/bit.cir
diff --git a/clang/include/clang/CIR/Dialect/IR/CIROps.td b/clang/include/clang/CIR/Dialect/IR/CIROps.td
index 694e3691c9361..8d5d34782c9a8 100644
--- a/clang/include/clang/CIR/Dialect/IR/CIROps.td
+++ b/clang/include/clang/CIR/Dialect/IR/CIROps.td
@@ -2773,6 +2773,8 @@ class CIR_BitOpBase<string mnemonic, TypeConstraint operandTy>
let assemblyFormat = [{
$input `:` type($result) attr-dict
}];
+
+ let hasFolder = 1;
}
class CIR_BitZeroCountOpBase<string mnemonic, TypeConstraint operandTy>
@@ -2980,6 +2982,8 @@ def CIR_RotateOp : CIR_Op<"rotate", [Pure, SameOperandsAndResultType]> {
bool isRotateLeft() { return getRotateLeft(); }
bool isRotateRight() { return !getRotateLeft(); }
}];
+
+ let hasFolder = 1;
}
//===----------------------------------------------------------------------===//
diff --git a/clang/include/clang/CIR/MissingFeatures.h b/clang/include/clang/CIR/MissingFeatures.h
index 37e0a4c8c1b6b..946cccaa6b721 100644
--- a/clang/include/clang/CIR/MissingFeatures.h
+++ b/clang/include/clang/CIR/MissingFeatures.h
@@ -284,6 +284,7 @@ struct MissingFeatures {
// Future CIR attributes
static bool optInfoAttr() { return false; }
+ static bool poisonAttr() { return false; }
};
} // namespace cir
diff --git a/clang/lib/CIR/Dialect/IR/CIRDialect.cpp b/clang/lib/CIR/Dialect/IR/CIRDialect.cpp
index f0416b6aba6e4..4c67873770895 100644
--- a/clang/lib/CIR/Dialect/IR/CIRDialect.cpp
+++ b/clang/lib/CIR/Dialect/IR/CIRDialect.cpp
@@ -17,6 +17,7 @@
#include "mlir/Interfaces/ControlFlowInterfaces.h"
#include "mlir/Interfaces/FunctionImplementation.h"
+#include "mlir/Support/LLVM.h"
#include "clang/CIR/Dialect/IR/CIROpsDialect.cpp.inc"
#include "clang/CIR/Dialect/IR/CIROpsEnums.cpp.inc"
@@ -2132,6 +2133,126 @@ LogicalResult cir::ComplexImagPtrOp::verify() {
return success();
}
+//===----------------------------------------------------------------------===//
+// Bit manipulation operations
+//===----------------------------------------------------------------------===//
+
+static OpFoldResult
+foldUnaryBitOp(mlir::Attribute inputAttr,
+ llvm::function_ref<llvm::APInt(const llvm::APInt &)> func,
+ bool poisonZero = false) {
+ auto input = mlir::dyn_cast_if_present<IntAttr>(inputAttr);
+ if (!input)
+ return nullptr;
+
+ llvm::APInt inputValue = input.getValue();
+ if (poisonZero && inputValue.isZero()) {
+ // TODO(cir): maybe we should return a poison value here?
+ assert(!MissingFeatures::poisonAttr());
+ return nullptr;
+ }
+
+ llvm::APInt resultValue = func(inputValue);
+ return IntAttr::get(input.getType(), resultValue);
+}
+
+OpFoldResult BitClrsbOp::fold(FoldAdaptor adaptor) {
+ return foldUnaryBitOp(adaptor.getInput(), [](const llvm::APInt &inputValue) {
+ unsigned resultValue =
+ inputValue.getBitWidth() - inputValue.getSignificantBits();
+ return llvm::APInt(inputValue.getBitWidth(), resultValue);
+ });
+}
+
+OpFoldResult BitClzOp::fold(FoldAdaptor adaptor) {
+ return foldUnaryBitOp(
+ adaptor.getInput(),
+ [](const llvm::APInt &inputValue) {
+ unsigned resultValue = inputValue.countLeadingZeros();
+ return llvm::APInt(inputValue.getBitWidth(), resultValue);
+ },
+ getPoisonZero());
+}
+
+OpFoldResult BitCtzOp::fold(FoldAdaptor adaptor) {
+ return foldUnaryBitOp(
+ adaptor.getInput(),
+ [](const llvm::APInt &inputValue) {
+ return llvm::APInt(inputValue.getBitWidth(),
+ inputValue.countTrailingZeros());
+ },
+ getPoisonZero());
+}
+
+OpFoldResult BitParityOp::fold(FoldAdaptor adaptor) {
+ return foldUnaryBitOp(adaptor.getInput(), [](const llvm::APInt &inputValue) {
+ return llvm::APInt(inputValue.getBitWidth(), inputValue.popcount() % 2);
+ });
+}
+
+OpFoldResult BitPopcountOp::fold(FoldAdaptor adaptor) {
+ return foldUnaryBitOp(adaptor.getInput(), [](const llvm::APInt &inputValue) {
+ return llvm::APInt(inputValue.getBitWidth(), inputValue.popcount());
+ });
+}
+
+OpFoldResult BitReverseOp::fold(FoldAdaptor adaptor) {
+ return foldUnaryBitOp(adaptor.getInput(), [](const llvm::APInt &inputValue) {
+ return inputValue.reverseBits();
+ });
+}
+
+OpFoldResult ByteSwapOp::fold(FoldAdaptor adaptor) {
+ return foldUnaryBitOp(adaptor.getInput(), [](const llvm::APInt &inputValue) {
+ return inputValue.byteSwap();
+ });
+}
+
+OpFoldResult RotateOp::fold(FoldAdaptor adaptor) {
+ auto input = mlir::dyn_cast_if_present<IntAttr>(adaptor.getInput());
+ auto amount = mlir::dyn_cast_if_present<IntAttr>(adaptor.getAmount());
+ if (!input && !amount)
+ return nullptr;
+
+ // We could fold cir.rotate even if one of its two operands is not a constant:
+ // - `cir.rotate left/right %0, 0` could be folded into just %0 even if %0
+ // is not a constant.
+ // - `cir.rotate left/right 0/0b111...111, %0` could be folded into 0 or
+ // 0b111...111 even if %0 is not a constant.
+
+ llvm::APInt inputValue;
+ if (input) {
+ inputValue = input.getValue();
+ if (inputValue.isZero() || inputValue.isAllOnes()) {
+ // An input value of all 0s or all 1s will not change after rotation
+ return input;
+ }
+ }
+
+ uint64_t amountValue;
+ if (amount) {
+ amountValue = amount.getValue().urem(getInput().getType().getWidth());
+ if (amountValue == 0) {
+ // A shift amount of 0 will not change the input value
+ return getInput();
+ }
+ }
+
+ if (!input || !amount)
+ return nullptr;
+
+ assert(inputValue.getBitWidth() == getInput().getType().getWidth() &&
+ "input value must have the same bit width as the input type");
+
+ llvm::APInt resultValue;
+ if (isRotateLeft())
+ resultValue = inputValue.rotl(amountValue);
+ else
+ resultValue = inputValue.rotr(amountValue);
+
+ return IntAttr::get(input.getContext(), input.getType(), resultValue);
+}
+
//===----------------------------------------------------------------------===//
// TableGen'd op method definitions
//===----------------------------------------------------------------------===//
diff --git a/clang/lib/CIR/Dialect/Transforms/CIRCanonicalize.cpp b/clang/lib/CIR/Dialect/Transforms/CIRCanonicalize.cpp
index e505db50d3609..2143f167ba2c8 100644
--- a/clang/lib/CIR/Dialect/Transforms/CIRCanonicalize.cpp
+++ b/clang/lib/CIR/Dialect/Transforms/CIRCanonicalize.cpp
@@ -143,7 +143,8 @@ void CIRCanonicalizePass::runOnOperation() {
if (isa<BrOp, BrCondOp, CastOp, ScopeOp, SwitchOp, SelectOp, UnaryOp,
ComplexCreateOp, ComplexImagOp, ComplexRealOp, VecCmpOp,
VecCreateOp, VecExtractOp, VecShuffleOp, VecShuffleDynamicOp,
- VecTernaryOp>(op))
+ VecTernaryOp, BitClrsbOp, BitClzOp, BitCtzOp, BitParityOp,
+ BitPopcountOp, BitReverseOp, ByteSwapOp, RotateOp>(op))
ops.push_back(op);
});
diff --git a/clang/test/CIR/Transforms/bit.cir b/clang/test/CIR/Transforms/bit.cir
new file mode 100644
index 0000000000000..c85b05de4d9ea
--- /dev/null
+++ b/clang/test/CIR/Transforms/bit.cir
@@ -0,0 +1,141 @@
+// RUN: cir-opt -cir-canonicalize -cir-simplify -o %t.cir %s
+// RUN: FileCheck --input-file=%t.cir %s
+
+!s32i = !cir.int<s, 32>
+!u32i = !cir.int<u, 32>
+
+module {
+ cir.func @fold_clrsb() -> !s32i {
+ %0 = cir.const #cir.int<1> : !s32i
+ %1 = cir.clrsb %0 : !s32i
+ cir.return %1 : !s32i
+ }
+ // CHECK-LABEL: @fold_clrsb
+ // CHECK-NEXT: %[[R:.+]] = cir.const #cir.int<30> : !s32i
+ // CHECK-NEXT: cir.return %[[R]] : !s32i
+ // CHECK-NEXT: }
+
+ cir.func @fold_clz() -> !u32i {
+ %0 = cir.const #cir.int<1> : !u32i
+ %1 = cir.clz %0 : !u32i
+ cir.return %1 : !u32i
+ }
+ // CHECK-LABEL: @fold_clz
+ // CHECK-NEXT: %[[R:.+]] = cir.const #cir.int<31> : !u32i
+ // CHECK-NEXT: cir.return %[[R]] : !u32i
+ // CHECK-NEXT: }
+
+ cir.func @fold_ctz() -> !u32i {
+ %0 = cir.const #cir.int<2> : !u32i
+ %1 = cir.ctz %0 : !u32i
+ cir.return %1 : !u32i
+ }
+ // CHECK-LABEL: @fold_ctz
+ // CHECK-NEXT: %[[R:.+]] = cir.const #cir.int<1> : !u32i
+ // CHECK-NEXT: cir.return %[[R]] : !u32i
+ // CHECK-NEXT: }
+
+ cir.func @fold_parity() -> !u32i {
+ // 0xdeadbeef is 0b1101_1110_1010_1101_1011_1110_1110_1111
+ // 0xdeadbeef contains 24 ones
+ %0 = cir.const #cir.int<0xdeadbeef> : !u32i
+ %1 = cir.parity %0 : !u32i
+ cir.return %1 : !u32i
+ }
+ // CHECK-LABEL: @fold_parity
+ // CHECK-NEXT: %[[R:.+]] = cir.const #cir.int<0> : !u32i
+ // CHECK-NEXT: cir.return %[[R]] : !u32i
+ // CHECK-NEXT: }
+
+ cir.func @fold_popcount() -> !u32i {
+ // 0xdeadbeef is 0b1101_1110_1010_1101_1011_1110_1110_1111
+ // 0xdeadbeef contains 24 ones
+ %0 = cir.const #cir.int<0xdeadbeef> : !u32i
+ %1 = cir.popcount %0 : !u32i
+ cir.return %1 : !u32i
+ }
+ // CHECK-LABEL: @fold_popcount
+ // CHECK-NEXT: %[[R:.+]] = cir.const #cir.int<24> : !u32i
+ // CHECK-NEXT: cir.return %[[R]] : !u32i
+ // CHECK-NEXT: }
+
+ cir.func @fold_bitreverse() -> !u32i {
+ // 0xdeadbeef is 0b1101_1110_1010_1101_1011_1110_1110_1111
+ %0 = cir.const #cir.int<0xdeadbeef> : !u32i
+ %1 = cir.bitreverse %0 : !u32i
+ cir.return %1 : !u32i
+ }
+ // CHECK-LABEL: @fold_bitreverse
+ // 4152210811 is 0b1111_0111_0111_1101_1011_0101_0111_1011
+ // CHECK-NEXT: %[[R:.+]] = cir.const #cir.int<4152210811> : !u32i
+ // CHECK-NEXT: cir.return %[[R]] : !u32i
+ // CHECK-NEXT: }
+
+ cir.func @fold_byte_swap() -> !u32i {
+ %0 = cir.const #cir.int<0xdeadbeef> : !u32i
+ %1 = cir.byte_swap %0 : !u32i
+ cir.return %1 : !u32i
+ }
+ // CHECK-LABEL: @fold_byte_swap
+ // 4022250974 is 0xefbeadde
+ // CHECK-NEXT: %[[R:.+]] = cir.const #cir.int<4022250974> : !u32i
+ // CHECK-NEXT: cir.return %[[R]] : !u32i
+ // CHECK-NEXT: }
+
+ cir.func @fold_rotate_input_all_zeros(%arg0 : !u32i) -> !u32i {
+ %0 = cir.const #cir.int<0> : !u32i
+ %1 = cir.rotate left %0, %arg0 : !u32i
+ cir.return %1 : !u32i
+ }
+ // CHECK-LABEL: @fold_rotate_input_all_zeros
+ // CHECK-NEXT: %[[R:.+]] = cir.const #cir.int<0> : !u32i
+ // CHECK-NEXT: cir.return %[[R]] : !u32i
+ // CHECK-NEXT: }
+
+ cir.func @fold_rotate_input_all_ones(%arg0 : !u32i) -> !u32i {
+ // 4294967295 is 0b1111_1111_1111_1111_1111_1111_1111_1111
+ %0 = cir.const #cir.int<4294967295> : !u32i
+ %1 = cir.rotate left %0, %arg0 : !u32i
+ cir.return %1 : !u32i
+ }
+ // CHECK-LABEL: @fold_rotate_input_all_ones
+ // CHECK-NEXT: %[[R:.+]] = cir.const #cir.int<4294967295> : !u32i
+ // CHECK-NEXT: cir.return %[[R]] : !u32i
+ // CHECK-NEXT: }
+
+ cir.func @fold_rotate_zero_amount(%arg0 : !u32i) -> !u32i {
+ %0 = cir.const #cir.int<32> : !u32i
+ %1 = cir.rotate left %arg0, %0 : !u32i
+ cir.return %1 : !u32i
+ }
+ // CHECK-LABEL: @fold_rotate_zero_amount
+ // CHECK-SAME: (%[[R:.+]]: !u32i)
+ // CHECK-NEXT: cir.return %[[R]] : !u32i
+ // CHECK-NEXT: }
+
+ cir.func @fold_rotate_left() -> !u32i {
+ // 0xdeadbeef is 0b1101_1110_1010_1101_1011_1110_1110_1111
+ %0 = cir.const #cir.int<0xdeadbeef> : !u32i
+ %1 = cir.const #cir.int<8> : !u32i
+ %2 = cir.rotate left %0, %1 : !u32i
+ cir.return %2 : !u32i
+ }
+ // CHECK-LABEL: @fold_rotate_left
+ // 2914971614 is 0b1010_1101_1011_1110_1110_1111_1101_1110
+ // CHECK-NEXT: %[[R:.+]] = cir.const #cir.int<2914971614> : !u32i
+ // CHECK-NEXT: cir.return %[[R]] : !u32i
+ // CHECK-NEXT: }
+
+ cir.func @fold_rotate_right() -> !u32i {
+ // 0xdeadbeef is 0b1101_1110_1010_1101_1011_1110_1110_1111
+ %0 = cir.const #cir.int<0xdeadbeef> : !u32i
+ %1 = cir.const #cir.int<8> : !u32i
+ %2 = cir.rotate right %0, %1 : !u32i
+ cir.return %2 : !u32i
+ }
+ // CHECK-LABEL: @fold_rotate_right
+ // 4260027374 is 0b1110_1111_1101_1110_1010_1101_1011_1110
+ // CHECK-NEXT: %[[R:.+]] = cir.const #cir.int<4024348094> : !u32i
+ // CHECK-NEXT: cir.return %[[R]] : !u32i
+ // CHECK-NEXT: }
+}
More information about the cfe-commits
mailing list