Add MCSymbolizer for symbolic/annotated disassembly.
This is a basic first step towards symbolization of disassembled
instructions. This used to be done using externally provided (C API)
callbacks. This patch introduces:
- the MCSymbolizer class, that mimics the same functions that were used
in the X86 and ARM disassemblers to symbolize immediate operands and
to annotate loads based off PC (for things like c string literals).
- the MCExternalSymbolizer class, which implements the old C API.
- the MCRelocationInfo class, which provides a way for targets to
translate relocations (either object::RelocationRef, or disassembler
C API VariantKinds) to MCExprs.
- the MCObjectSymbolizer class, which does symbolization using what it
finds in an object::ObjectFile. This makes simple symbolization (with
no fancy relocation stuff) work for all object formats!
- x86-64 Mach-O and ELF MCRelocationInfos.
- A basic ARM Mach-O MCRelocationInfo, that provides just enough to
support the C API VariantKinds.
Most of what works in otool (the only user of the old symbolization API
that I know of) for x86-64 symbolic disassembly (-tvV) works, namely:
- symbol references: call _foo; jmp 15 <_foo+50>
- relocations: call _foo-_bar; call _foo-4
- __cf?string: leaq 193(%rip), %rax ## literal pool for "hello"
Stub support is the main missing part (because libObject doesn't know,
among other things, about mach-o indirect symbols).
As for the MCSymbolizer API, instead of relying on the disassemblers
to call the tryAdding* methods, maybe this could be done automagically
using InstrInfo? For instance, even though PC-relative LEAs are used
to get the address of string literals in a typical Mach-O file, a MOV
would be used in an ELF file. And right now, the explicit symbolization
only recognizes PC-relative LEAs. InstrInfo should have already have
most of what is needed to know what to symbolize, so this can
definitely be improved.
I'd also like to remove object::RelocationRef::getValueString (it seems
only used by relocation printing in objdump), as simply printing the
created MCExpr is definitely enough (and cleaner than string concats).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@182625 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-24 00:39:57 +00:00
|
|
|
//===-- llvm/MC/MCSymbolizer.h - MCSymbolizer class -------------*- C++ -*-===//
|
|
|
|
//
|
|
|
|
// The LLVM Compiler Infrastructure
|
|
|
|
//
|
|
|
|
// This file is distributed under the University of Illinois Open Source
|
|
|
|
// License. See LICENSE.TXT for details.
|
|
|
|
//
|
|
|
|
//===----------------------------------------------------------------------===//
|
|
|
|
//
|
|
|
|
// This file contains the declaration of the MCSymbolizer class, which is used
|
|
|
|
// to symbolize instructions decoded from an object, that is, transform their
|
|
|
|
// immediate operands to MCExprs.
|
|
|
|
//
|
|
|
|
//===----------------------------------------------------------------------===//
|
|
|
|
|
|
|
|
#ifndef LLVM_MC_MCSYMBOLIZER_H
|
|
|
|
#define LLVM_MC_MCSYMBOLIZER_H
|
|
|
|
|
|
|
|
#include "llvm/MC/MCRelocationInfo.h"
|
|
|
|
#include "llvm/Support/Compiler.h"
|
|
|
|
#include "llvm/Support/DataTypes.h"
|
2014-03-27 02:58:32 +00:00
|
|
|
#include <cassert>
|
2014-03-06 05:51:42 +00:00
|
|
|
#include <memory>
|
Add MCSymbolizer for symbolic/annotated disassembly.
This is a basic first step towards symbolization of disassembled
instructions. This used to be done using externally provided (C API)
callbacks. This patch introduces:
- the MCSymbolizer class, that mimics the same functions that were used
in the X86 and ARM disassemblers to symbolize immediate operands and
to annotate loads based off PC (for things like c string literals).
- the MCExternalSymbolizer class, which implements the old C API.
- the MCRelocationInfo class, which provides a way for targets to
translate relocations (either object::RelocationRef, or disassembler
C API VariantKinds) to MCExprs.
- the MCObjectSymbolizer class, which does symbolization using what it
finds in an object::ObjectFile. This makes simple symbolization (with
no fancy relocation stuff) work for all object formats!
- x86-64 Mach-O and ELF MCRelocationInfos.
- A basic ARM Mach-O MCRelocationInfo, that provides just enough to
support the C API VariantKinds.
Most of what works in otool (the only user of the old symbolization API
that I know of) for x86-64 symbolic disassembly (-tvV) works, namely:
- symbol references: call _foo; jmp 15 <_foo+50>
- relocations: call _foo-_bar; call _foo-4
- __cf?string: leaq 193(%rip), %rax ## literal pool for "hello"
Stub support is the main missing part (because libObject doesn't know,
among other things, about mach-o indirect symbols).
As for the MCSymbolizer API, instead of relying on the disassemblers
to call the tryAdding* methods, maybe this could be done automagically
using InstrInfo? For instance, even though PC-relative LEAs are used
to get the address of string literals in a typical Mach-O file, a MOV
would be used in an ELF file. And right now, the explicit symbolization
only recognizes PC-relative LEAs. InstrInfo should have already have
most of what is needed to know what to symbolize, so this can
definitely be improved.
I'd also like to remove object::RelocationRef::getValueString (it seems
only used by relocation printing in objdump), as simply printing the
created MCExpr is definitely enough (and cleaner than string concats).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@182625 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-24 00:39:57 +00:00
|
|
|
|
|
|
|
namespace llvm {
|
|
|
|
|
|
|
|
class MCContext;
|
|
|
|
class MCInst;
|
|
|
|
class raw_ostream;
|
|
|
|
|
|
|
|
/// \brief Symbolize and annotate disassembled instructions.
|
|
|
|
///
|
|
|
|
/// For now this mimics the old symbolization logic (from both ARM and x86), that
|
|
|
|
/// relied on user-provided (C API) callbacks to do the actual symbol lookup in
|
|
|
|
/// the object file. This was moved to MCExternalSymbolizer.
|
|
|
|
/// A better API would not rely on actually calling the two methods here from
|
|
|
|
/// inside each disassembler, but would use the instr info to determine what
|
|
|
|
/// operands are actually symbolizable, and in what way. I don't think this
|
|
|
|
/// information exists right now.
|
|
|
|
class MCSymbolizer {
|
|
|
|
MCSymbolizer(const MCSymbolizer &) LLVM_DELETED_FUNCTION;
|
|
|
|
void operator=(const MCSymbolizer &) LLVM_DELETED_FUNCTION;
|
|
|
|
|
|
|
|
protected:
|
|
|
|
MCContext &Ctx;
|
2014-03-06 05:51:42 +00:00
|
|
|
std::unique_ptr<MCRelocationInfo> RelInfo;
|
Add MCSymbolizer for symbolic/annotated disassembly.
This is a basic first step towards symbolization of disassembled
instructions. This used to be done using externally provided (C API)
callbacks. This patch introduces:
- the MCSymbolizer class, that mimics the same functions that were used
in the X86 and ARM disassemblers to symbolize immediate operands and
to annotate loads based off PC (for things like c string literals).
- the MCExternalSymbolizer class, which implements the old C API.
- the MCRelocationInfo class, which provides a way for targets to
translate relocations (either object::RelocationRef, or disassembler
C API VariantKinds) to MCExprs.
- the MCObjectSymbolizer class, which does symbolization using what it
finds in an object::ObjectFile. This makes simple symbolization (with
no fancy relocation stuff) work for all object formats!
- x86-64 Mach-O and ELF MCRelocationInfos.
- A basic ARM Mach-O MCRelocationInfo, that provides just enough to
support the C API VariantKinds.
Most of what works in otool (the only user of the old symbolization API
that I know of) for x86-64 symbolic disassembly (-tvV) works, namely:
- symbol references: call _foo; jmp 15 <_foo+50>
- relocations: call _foo-_bar; call _foo-4
- __cf?string: leaq 193(%rip), %rax ## literal pool for "hello"
Stub support is the main missing part (because libObject doesn't know,
among other things, about mach-o indirect symbols).
As for the MCSymbolizer API, instead of relying on the disassemblers
to call the tryAdding* methods, maybe this could be done automagically
using InstrInfo? For instance, even though PC-relative LEAs are used
to get the address of string literals in a typical Mach-O file, a MOV
would be used in an ELF file. And right now, the explicit symbolization
only recognizes PC-relative LEAs. InstrInfo should have already have
most of what is needed to know what to symbolize, so this can
definitely be improved.
I'd also like to remove object::RelocationRef::getValueString (it seems
only used by relocation printing in objdump), as simply printing the
created MCExpr is definitely enough (and cleaner than string concats).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@182625 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-24 00:39:57 +00:00
|
|
|
|
|
|
|
public:
|
|
|
|
/// \brief Construct an MCSymbolizer, taking ownership of \p RelInfo.
|
2014-03-27 02:42:52 +00:00
|
|
|
MCSymbolizer(MCContext &Ctx, std::unique_ptr<MCRelocationInfo> RelInfo)
|
2014-03-27 02:49:18 +00:00
|
|
|
: Ctx(Ctx), RelInfo(std::move(RelInfo)) {
|
|
|
|
}
|
2014-03-27 02:42:52 +00:00
|
|
|
|
Add MCSymbolizer for symbolic/annotated disassembly.
This is a basic first step towards symbolization of disassembled
instructions. This used to be done using externally provided (C API)
callbacks. This patch introduces:
- the MCSymbolizer class, that mimics the same functions that were used
in the X86 and ARM disassemblers to symbolize immediate operands and
to annotate loads based off PC (for things like c string literals).
- the MCExternalSymbolizer class, which implements the old C API.
- the MCRelocationInfo class, which provides a way for targets to
translate relocations (either object::RelocationRef, or disassembler
C API VariantKinds) to MCExprs.
- the MCObjectSymbolizer class, which does symbolization using what it
finds in an object::ObjectFile. This makes simple symbolization (with
no fancy relocation stuff) work for all object formats!
- x86-64 Mach-O and ELF MCRelocationInfos.
- A basic ARM Mach-O MCRelocationInfo, that provides just enough to
support the C API VariantKinds.
Most of what works in otool (the only user of the old symbolization API
that I know of) for x86-64 symbolic disassembly (-tvV) works, namely:
- symbol references: call _foo; jmp 15 <_foo+50>
- relocations: call _foo-_bar; call _foo-4
- __cf?string: leaq 193(%rip), %rax ## literal pool for "hello"
Stub support is the main missing part (because libObject doesn't know,
among other things, about mach-o indirect symbols).
As for the MCSymbolizer API, instead of relying on the disassemblers
to call the tryAdding* methods, maybe this could be done automagically
using InstrInfo? For instance, even though PC-relative LEAs are used
to get the address of string literals in a typical Mach-O file, a MOV
would be used in an ELF file. And right now, the explicit symbolization
only recognizes PC-relative LEAs. InstrInfo should have already have
most of what is needed to know what to symbolize, so this can
definitely be improved.
I'd also like to remove object::RelocationRef::getValueString (it seems
only used by relocation printing in objdump), as simply printing the
created MCExpr is definitely enough (and cleaner than string concats).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@182625 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-24 00:39:57 +00:00
|
|
|
virtual ~MCSymbolizer();
|
|
|
|
|
|
|
|
/// \brief Try to add a symbolic operand instead of \p Value to the MCInst.
|
|
|
|
///
|
|
|
|
/// Instead of having a difficult to read immediate, a symbolic operand would
|
|
|
|
/// represent this immediate in a more understandable way, for instance as a
|
|
|
|
/// symbol or an offset from a symbol. Relocations can also be used to enrich
|
|
|
|
/// the symbolic expression.
|
|
|
|
/// @param Inst - The MCInst where to insert the symbolic operand.
|
|
|
|
/// @param cStream - Stream to print comments and annotations on.
|
|
|
|
/// @param Value - Operand value, pc-adjusted by the caller if necessary.
|
|
|
|
/// @param Address - Load address of the instruction.
|
|
|
|
/// @param IsBranch - Is the instruction a branch?
|
|
|
|
/// @param Offset - Byte offset of the operand inside the inst.
|
|
|
|
/// @param InstSize - Size of the instruction in bytes.
|
|
|
|
/// @return Whether a symbolic operand was added.
|
|
|
|
virtual bool tryAddingSymbolicOperand(MCInst &Inst, raw_ostream &cStream,
|
|
|
|
int64_t Value, uint64_t Address,
|
|
|
|
bool IsBranch, uint64_t Offset,
|
|
|
|
uint64_t InstSize) = 0;
|
|
|
|
|
|
|
|
/// \brief Try to add a comment on the PC-relative load.
|
|
|
|
/// For instance, in Mach-O, this is used to add annotations to instructions
|
|
|
|
/// that use C string literals, as found in __cstring.
|
|
|
|
virtual void tryAddingPcLoadReferenceComment(raw_ostream &cStream,
|
|
|
|
int64_t Value,
|
|
|
|
uint64_t Address) = 0;
|
|
|
|
};
|
|
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
#endif
|