Files
revng-revng/lib/CliftTransforms/PointerArithmetic.h
Andrea Gussoni 02cc6923d0 EmitFieldAccesses: implement pass
The `EmitFieldAccesses` pass transforms `clift` by taking pointer-typed
expressions computed via integerr arithmetic with type-safe field
accesses and array accesses.

The transformation is split in three main phases:
1) `PointerArithmetic` computation.
2) `BestTraversal` computation.
3) `FieldAccess` `clift` rewrite.

The high level driver is implemented in the `EmitFieldAccesses` header
and cpp, while the nested 3 phases are implemented respectively in
`PointerArithmetic`, `BestTraversal` and `FieldAccessReplacement`.

The `computerPointerArithmetic` phase is concerned with taking a
pointer-typed `ExpressionOp`, called `PointerToReplace`, and expressing
it in a `BasePointer+Offset` form.

The `computeBestTraversal` phase is concerned with computing the best
traversal of the type pointed to by `BasePointer`, that can be used to
rewrite the pointer arithmetic in `clift` with just field accesses and
array subscripts.

The `replaceFieldAccess` phase takes the `Traversal` computed at the
previous step, and actually rewrites in `clift` the `PointerToReplace`
in terms of field accesses and array accesses w.r.t. the `BasePointer`.
2026-03-18 17:25:13 +01:00

79 lines
2.2 KiB
C++

#pragma once
//
// This file is distributed under the MIT License. See LICENSE.md for details.
//
#include <optional>
#include "llvm/ADT/APInt.h"
#include "llvm/ADT/SmallVector.h"
#include "mlir/IR/Value.h"
#include "revng/Clift/Clift.h"
#include "revng/Support/Debug.h"
/// Represents a pointer-typed expression decomposed into a `BasePointer` and an
/// `Offset` expression. The offset is a constant `BaseOffset`, plus a linear
/// combination of strided terms, which capture array index patterns
struct PointerArithmetic {
/// Represents an `Index` with both a variable and a constant component,
/// e.g., `array[i+4]`
struct Index {
mlir::Value Variable;
llvm::APInt Constant;
};
/// Represents a strided term in the `PointerArithmetic`
struct StridedTerm {
llvm::APInt Stride;
Index Idx;
StridedTerm(llvm::APInt Stride, Index Idx) :
Stride(std::move(Stride)), Idx(std::move(Idx)) {}
void dump() const debug_function;
};
/// Represents the offset with possible strided terms in the
/// `PointerArithmetic`
struct OffsetExpression {
/// Represents the constant part of the `Offset`
llvm::APInt BaseOffset;
/// Holds the terms of the linear combination components of the `Offset`
llvm::SmallVector<StridedTerm> LinearCombination;
OffsetExpression(unsigned BitWidth);
OffsetExpression(llvm::APInt Offset);
void dump() const debug_function;
};
/// The base pointer the `PointerArithmetic` is expressed relative to
/// `BasePointer` object, which is where the root of the computation lies
mlir::Value BasePointer;
/// The offset w.r.t. the `BasePointer`
OffsetExpression Offset;
/// We define a `PointerArithmetic` with an empty `BasePointer` a _numeric_
bool isNumeric() const;
/// Check if this is an address (has base pointer) arithmetic
bool isAddress() const;
/// Verify the invariants for the strides contained in the PointerArithmetic
bool verify() const;
/// Dump method
void dump() const debug_function;
};
/// `PointerArithmetic` computation function entrypoint
std::optional<PointerArithmetic>
computePointerArithmetic(mlir::clift::ExpressionOpInterface PointerToReplace);