Summary: This is necessary to support parsing expressions like ".cfa -16 + ^", as that format is used in breakpad STACK CFI expressions. Since the PDB expressions use the same parser, this change will affect them too, but I don't believe that should be a problem in practice. If PDBs do contain the negative values, it's very likely that they are intended to be parsed the same way, and if they don't, then it doesn't matter. In case that we do ever need to handle this differently, we can always make the parser behavior customizable, or just use a different parser. To make sure that the integer size is big enough for everyone, I switch from using a (unsigned) 32-bit integer to a 64-bit (signed) one. Reviewers: amccarth, clayborg, aleksandr.urakov Subscribers: markmentovai, lldb-commits Differential Revision: https://reviews.llvm.org/D61311 llvm-svn: 360166
228 lines
6.3 KiB
C++
228 lines
6.3 KiB
C++
//===-- PostfixExpression.cpp -----------------------------------*- C++ -*-===//
|
|
//
|
|
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
|
|
// See https://llvm.org/LICENSE.txt for license information.
|
|
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
|
|
//
|
|
//===----------------------------------------------------------------------===//
|
|
//
|
|
// This file implements support for postfix expressions found in several symbol
|
|
// file formats, and their conversion to DWARF.
|
|
//
|
|
//===----------------------------------------------------------------------===//
|
|
|
|
#include "lldb/Symbol/PostfixExpression.h"
|
|
#include "lldb/Core/dwarf.h"
|
|
#include "lldb/Utility/Stream.h"
|
|
#include "llvm/ADT/StringExtras.h"
|
|
|
|
using namespace lldb_private;
|
|
using namespace lldb_private::postfix;
|
|
|
|
static llvm::Optional<BinaryOpNode::OpType>
|
|
GetBinaryOpType(llvm::StringRef token) {
|
|
if (token.size() != 1)
|
|
return llvm::None;
|
|
switch (token[0]) {
|
|
case '@':
|
|
return BinaryOpNode::Align;
|
|
case '-':
|
|
return BinaryOpNode::Minus;
|
|
case '+':
|
|
return BinaryOpNode::Plus;
|
|
}
|
|
return llvm::None;
|
|
}
|
|
|
|
static llvm::Optional<UnaryOpNode::OpType>
|
|
GetUnaryOpType(llvm::StringRef token) {
|
|
if (token == "^")
|
|
return UnaryOpNode::Deref;
|
|
return llvm::None;
|
|
}
|
|
|
|
Node *postfix::Parse(llvm::StringRef expr, llvm::BumpPtrAllocator &alloc) {
|
|
llvm::SmallVector<Node *, 4> stack;
|
|
|
|
llvm::StringRef token;
|
|
while (std::tie(token, expr) = getToken(expr), !token.empty()) {
|
|
if (auto op_type = GetBinaryOpType(token)) {
|
|
// token is binary operator
|
|
if (stack.size() < 2)
|
|
return nullptr;
|
|
|
|
Node *right = stack.pop_back_val();
|
|
Node *left = stack.pop_back_val();
|
|
stack.push_back(MakeNode<BinaryOpNode>(alloc, *op_type, *left, *right));
|
|
continue;
|
|
}
|
|
|
|
if (auto op_type = GetUnaryOpType(token)) {
|
|
// token is unary operator
|
|
if (stack.empty())
|
|
return nullptr;
|
|
|
|
Node *operand = stack.pop_back_val();
|
|
stack.push_back(MakeNode<UnaryOpNode>(alloc, *op_type, *operand));
|
|
continue;
|
|
}
|
|
|
|
int64_t value;
|
|
if (to_integer(token, value, 10)) {
|
|
// token is integer literal
|
|
stack.push_back(MakeNode<IntegerNode>(alloc, value));
|
|
continue;
|
|
}
|
|
|
|
stack.push_back(MakeNode<SymbolNode>(alloc, token));
|
|
}
|
|
|
|
if (stack.size() != 1)
|
|
return nullptr;
|
|
|
|
return stack.back();
|
|
}
|
|
|
|
namespace {
|
|
class SymbolResolver : public Visitor<bool> {
|
|
public:
|
|
SymbolResolver(llvm::function_ref<Node *(SymbolNode &symbol)> replacer)
|
|
: m_replacer(replacer) {}
|
|
|
|
using Visitor<bool>::Dispatch;
|
|
|
|
private:
|
|
bool Visit(BinaryOpNode &binary, Node *&) override {
|
|
return Dispatch(binary.Left()) && Dispatch(binary.Right());
|
|
}
|
|
|
|
bool Visit(InitialValueNode &, Node *&) override { return true; }
|
|
bool Visit(IntegerNode &, Node *&) override { return true; }
|
|
bool Visit(RegisterNode &, Node *&) override { return true; }
|
|
|
|
bool Visit(SymbolNode &symbol, Node *&ref) override {
|
|
if (Node *replacement = m_replacer(symbol)) {
|
|
ref = replacement;
|
|
if (replacement != &symbol)
|
|
return Dispatch(ref);
|
|
return true;
|
|
}
|
|
return false;
|
|
}
|
|
|
|
bool Visit(UnaryOpNode &unary, Node *&) override {
|
|
return Dispatch(unary.Operand());
|
|
}
|
|
|
|
llvm::function_ref<Node *(SymbolNode &symbol)> m_replacer;
|
|
};
|
|
|
|
class DWARFCodegen : public Visitor<> {
|
|
public:
|
|
DWARFCodegen(Stream &stream) : m_out_stream(stream) {}
|
|
|
|
using Visitor<>::Dispatch;
|
|
|
|
private:
|
|
void Visit(BinaryOpNode &binary, Node *&) override;
|
|
|
|
void Visit(InitialValueNode &val, Node *&) override;
|
|
|
|
void Visit(IntegerNode &integer, Node *&) override {
|
|
m_out_stream.PutHex8(DW_OP_consts);
|
|
m_out_stream.PutSLEB128(integer.GetValue());
|
|
++m_stack_depth;
|
|
}
|
|
|
|
void Visit(RegisterNode ®, Node *&) override;
|
|
|
|
void Visit(SymbolNode &symbol, Node *&) override {
|
|
llvm_unreachable("Symbols should have been resolved by now!");
|
|
}
|
|
|
|
void Visit(UnaryOpNode &unary, Node *&) override;
|
|
|
|
Stream &m_out_stream;
|
|
|
|
/// The number keeping track of the evaluation stack depth at any given
|
|
/// moment. Used for implementing InitialValueNodes. We start with
|
|
/// m_stack_depth = 1, assuming that the initial value is already on the
|
|
/// stack. This initial value will be the value of all InitialValueNodes. If
|
|
/// the expression does not contain InitialValueNodes, then m_stack_depth is
|
|
/// not used, and the generated expression will run correctly even without an
|
|
/// initial value.
|
|
size_t m_stack_depth = 1;
|
|
};
|
|
} // namespace
|
|
|
|
void DWARFCodegen::Visit(BinaryOpNode &binary, Node *&) {
|
|
Dispatch(binary.Left());
|
|
Dispatch(binary.Right());
|
|
|
|
switch (binary.GetOpType()) {
|
|
case BinaryOpNode::Plus:
|
|
m_out_stream.PutHex8(DW_OP_plus);
|
|
// NOTE: can be optimized by using DW_OP_plus_uconst opcpode
|
|
// if right child node is constant value
|
|
break;
|
|
case BinaryOpNode::Minus:
|
|
m_out_stream.PutHex8(DW_OP_minus);
|
|
break;
|
|
case BinaryOpNode::Align:
|
|
// emit align operator a @ b as
|
|
// a & ~(b - 1)
|
|
// NOTE: implicitly assuming that b is power of 2
|
|
m_out_stream.PutHex8(DW_OP_lit1);
|
|
m_out_stream.PutHex8(DW_OP_minus);
|
|
m_out_stream.PutHex8(DW_OP_not);
|
|
|
|
m_out_stream.PutHex8(DW_OP_and);
|
|
break;
|
|
}
|
|
--m_stack_depth; // Two pops, one push.
|
|
}
|
|
|
|
void DWARFCodegen::Visit(InitialValueNode &, Node *&) {
|
|
// We never go below the initial stack, so we can pick the initial value from
|
|
// the bottom of the stack at any moment.
|
|
assert(m_stack_depth >= 1);
|
|
m_out_stream.PutHex8(DW_OP_pick);
|
|
m_out_stream.PutHex8(m_stack_depth - 1);
|
|
++m_stack_depth;
|
|
}
|
|
|
|
void DWARFCodegen::Visit(RegisterNode ®, Node *&) {
|
|
uint32_t reg_num = reg.GetRegNum();
|
|
assert(reg_num != LLDB_INVALID_REGNUM);
|
|
|
|
if (reg_num > 31) {
|
|
m_out_stream.PutHex8(DW_OP_bregx);
|
|
m_out_stream.PutULEB128(reg_num);
|
|
} else
|
|
m_out_stream.PutHex8(DW_OP_breg0 + reg_num);
|
|
|
|
m_out_stream.PutSLEB128(0);
|
|
++m_stack_depth;
|
|
}
|
|
|
|
void DWARFCodegen::Visit(UnaryOpNode &unary, Node *&) {
|
|
Dispatch(unary.Operand());
|
|
|
|
switch (unary.GetOpType()) {
|
|
case UnaryOpNode::Deref:
|
|
m_out_stream.PutHex8(DW_OP_deref);
|
|
break;
|
|
}
|
|
// Stack depth unchanged.
|
|
}
|
|
|
|
bool postfix::ResolveSymbols(
|
|
Node *&node, llvm::function_ref<Node *(SymbolNode &)> replacer) {
|
|
return SymbolResolver(replacer).Dispatch(node);
|
|
}
|
|
|
|
void postfix::ToDWARF(Node &node, Stream &stream) {
|
|
Node *ptr = &node;
|
|
DWARFCodegen(stream).Dispatch(ptr);
|
|
}
|