2011-03-21 22:15:52 +00:00
|
|
|
//===-- RuntimeDyld.h - Run-time dynamic linker for MC-JIT ------*- C++ -*-===//
|
|
|
|
//
|
|
|
|
// The LLVM Compiler Infrastructure
|
|
|
|
//
|
|
|
|
// This file is distributed under the University of Illinois Open Source
|
|
|
|
// License. See LICENSE.TXT for details.
|
|
|
|
//
|
|
|
|
//===----------------------------------------------------------------------===//
|
|
|
|
//
|
|
|
|
// Implementation of the MC-JIT runtime dynamic linker.
|
|
|
|
//
|
|
|
|
//===----------------------------------------------------------------------===//
|
|
|
|
|
2011-03-23 19:52:00 +00:00
|
|
|
#define DEBUG_TYPE "dyld"
|
2011-03-21 22:15:52 +00:00
|
|
|
#include "llvm/ADT/OwningPtr.h"
|
2011-03-23 19:52:00 +00:00
|
|
|
#include "llvm/ADT/SmallVector.h"
|
2011-03-21 22:15:52 +00:00
|
|
|
#include "llvm/ADT/StringMap.h"
|
|
|
|
#include "llvm/ADT/StringRef.h"
|
2011-04-06 01:11:05 +00:00
|
|
|
#include "llvm/ADT/STLExtras.h"
|
2011-03-21 22:15:52 +00:00
|
|
|
#include "llvm/ADT/Twine.h"
|
|
|
|
#include "llvm/ExecutionEngine/RuntimeDyld.h"
|
|
|
|
#include "llvm/Object/MachOObject.h"
|
2011-03-23 19:52:00 +00:00
|
|
|
#include "llvm/Support/Debug.h"
|
|
|
|
#include "llvm/Support/ErrorHandling.h"
|
|
|
|
#include "llvm/Support/Format.h"
|
2011-03-21 22:15:52 +00:00
|
|
|
#include "llvm/Support/Memory.h"
|
|
|
|
#include "llvm/Support/MemoryBuffer.h"
|
|
|
|
#include "llvm/Support/system_error.h"
|
2011-03-23 19:52:00 +00:00
|
|
|
#include "llvm/Support/raw_ostream.h"
|
2011-03-21 22:15:52 +00:00
|
|
|
using namespace llvm;
|
|
|
|
using namespace llvm::object;
|
|
|
|
|
2011-04-05 23:54:31 +00:00
|
|
|
// Empty out-of-line virtual destructor as the key function.
|
|
|
|
RTDyldMemoryManager::~RTDyldMemoryManager() {}
|
|
|
|
|
2011-03-21 22:15:52 +00:00
|
|
|
namespace llvm {
|
|
|
|
class RuntimeDyldImpl {
|
2011-03-23 22:06:06 +00:00
|
|
|
unsigned CPUType;
|
|
|
|
unsigned CPUSubtype;
|
|
|
|
|
2011-04-04 23:04:39 +00:00
|
|
|
// The MemoryManager to load objects into.
|
|
|
|
RTDyldMemoryManager *MemMgr;
|
2011-03-29 21:03:05 +00:00
|
|
|
|
MCJIT lazy relocation resolution and symbol address re-assignment.
Add handling for tracking the relocations on symbols and resolving them.
Keep track of the relocations even after they are resolved so that if
the RuntimeDyld client moves the object, it can update the address and any
relocations to that object will be updated.
For our trival object file load/run test harness (llvm-rtdyld), this enables
relocations between functions located in the same object module. It should
be trivially extendable to load multiple objects with mutual references.
As a simple example, the following now works (running on x86_64 Darwin 10.6):
$ cat t.c
int bar() {
return 65;
}
int main() {
return bar();
}
$ clang t.c -fno-asynchronous-unwind-tables -o t.o -c
$ otool -vt t.o
t.o:
(__TEXT,__text) section
_bar:
0000000000000000 pushq %rbp
0000000000000001 movq %rsp,%rbp
0000000000000004 movl $0x00000041,%eax
0000000000000009 popq %rbp
000000000000000a ret
000000000000000b nopl 0x00(%rax,%rax)
_main:
0000000000000010 pushq %rbp
0000000000000011 movq %rsp,%rbp
0000000000000014 subq $0x10,%rsp
0000000000000018 movl $0x00000000,0xfc(%rbp)
000000000000001f callq 0x00000024
0000000000000024 addq $0x10,%rsp
0000000000000028 popq %rbp
0000000000000029 ret
$ llvm-rtdyld t.o -debug-only=dyld ; echo $?
Function sym: '_bar' @ 0
Function sym: '_main' @ 16
Extracting function: _bar from [0, 15]
allocated to 0x100153000
Extracting function: _main from [16, 41]
allocated to 0x100154000
Relocation at '_main' + 16 from '_bar(Word1: 0x2d000000)
Resolving relocation at '_main' + 16 (0x100154010) from '_bar (0x100153000)(pcrel, type: 2, Size: 4).
loaded '_main' at: 0x100154000
65
$
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@129388 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-12 21:20:41 +00:00
|
|
|
// FIXME: This all assumes we're dealing with external symbols for anything
|
|
|
|
// explicitly referenced. I.e., we can index by name and things
|
|
|
|
// will work out. In practice, this may not be the case, so we
|
|
|
|
// should find a way to effectively generalize.
|
2011-04-06 01:11:05 +00:00
|
|
|
|
|
|
|
// For each function, we have a MemoryBlock of it's instruction data.
|
|
|
|
StringMap<sys::MemoryBlock> Functions;
|
|
|
|
|
2011-03-21 22:15:52 +00:00
|
|
|
// Master symbol table. As modules are loaded and external symbols are
|
|
|
|
// resolved, their addresses are stored here.
|
MCJIT lazy relocation resolution and symbol address re-assignment.
Add handling for tracking the relocations on symbols and resolving them.
Keep track of the relocations even after they are resolved so that if
the RuntimeDyld client moves the object, it can update the address and any
relocations to that object will be updated.
For our trival object file load/run test harness (llvm-rtdyld), this enables
relocations between functions located in the same object module. It should
be trivially extendable to load multiple objects with mutual references.
As a simple example, the following now works (running on x86_64 Darwin 10.6):
$ cat t.c
int bar() {
return 65;
}
int main() {
return bar();
}
$ clang t.c -fno-asynchronous-unwind-tables -o t.o -c
$ otool -vt t.o
t.o:
(__TEXT,__text) section
_bar:
0000000000000000 pushq %rbp
0000000000000001 movq %rsp,%rbp
0000000000000004 movl $0x00000041,%eax
0000000000000009 popq %rbp
000000000000000a ret
000000000000000b nopl 0x00(%rax,%rax)
_main:
0000000000000010 pushq %rbp
0000000000000011 movq %rsp,%rbp
0000000000000014 subq $0x10,%rsp
0000000000000018 movl $0x00000000,0xfc(%rbp)
000000000000001f callq 0x00000024
0000000000000024 addq $0x10,%rsp
0000000000000028 popq %rbp
0000000000000029 ret
$ llvm-rtdyld t.o -debug-only=dyld ; echo $?
Function sym: '_bar' @ 0
Function sym: '_main' @ 16
Extracting function: _bar from [0, 15]
allocated to 0x100153000
Extracting function: _main from [16, 41]
allocated to 0x100154000
Relocation at '_main' + 16 from '_bar(Word1: 0x2d000000)
Resolving relocation at '_main' + 16 (0x100154010) from '_bar (0x100153000)(pcrel, type: 2, Size: 4).
loaded '_main' at: 0x100154000
65
$
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@129388 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-12 21:20:41 +00:00
|
|
|
StringMap<uint8_t*> SymbolTable;
|
|
|
|
|
|
|
|
// For each symbol, keep a list of relocations based on it. Anytime
|
|
|
|
// its address is reassigned (the JIT re-compiled the function, e.g.),
|
|
|
|
// the relocations get re-resolved.
|
|
|
|
struct RelocationEntry {
|
|
|
|
std::string Target; // Object this relocation is contained in.
|
|
|
|
uint64_t Offset; // Offset into the object for the relocation.
|
|
|
|
uint32_t Data; // Second word of the raw macho relocation entry.
|
|
|
|
int64_t Addend; // Addend encoded in the instruction itself, if any.
|
|
|
|
bool isResolved; // Has this relocation been resolved previously?
|
|
|
|
|
|
|
|
RelocationEntry(StringRef t, uint64_t offset, uint32_t data, int64_t addend)
|
|
|
|
: Target(t), Offset(offset), Data(data), Addend(addend),
|
|
|
|
isResolved(false) {}
|
|
|
|
};
|
|
|
|
typedef SmallVector<RelocationEntry, 4> RelocationList;
|
|
|
|
StringMap<RelocationList> Relocations;
|
|
|
|
|
|
|
|
// FIXME: Also keep a map of all the relocations contained in an object. Use
|
|
|
|
// this to dynamically answer whether all of the relocations in it have
|
|
|
|
// been resolved or not.
|
2011-03-21 22:15:52 +00:00
|
|
|
|
|
|
|
bool HasError;
|
|
|
|
std::string ErrorStr;
|
|
|
|
|
|
|
|
// Set the error state and record an error string.
|
|
|
|
bool Error(const Twine &Msg) {
|
|
|
|
ErrorStr = Msg.str();
|
|
|
|
HasError = true;
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
|
2011-04-06 01:11:05 +00:00
|
|
|
void extractFunction(StringRef Name, uint8_t *StartAddress,
|
|
|
|
uint8_t *EndAddress);
|
MCJIT lazy relocation resolution and symbol address re-assignment.
Add handling for tracking the relocations on symbols and resolving them.
Keep track of the relocations even after they are resolved so that if
the RuntimeDyld client moves the object, it can update the address and any
relocations to that object will be updated.
For our trival object file load/run test harness (llvm-rtdyld), this enables
relocations between functions located in the same object module. It should
be trivially extendable to load multiple objects with mutual references.
As a simple example, the following now works (running on x86_64 Darwin 10.6):
$ cat t.c
int bar() {
return 65;
}
int main() {
return bar();
}
$ clang t.c -fno-asynchronous-unwind-tables -o t.o -c
$ otool -vt t.o
t.o:
(__TEXT,__text) section
_bar:
0000000000000000 pushq %rbp
0000000000000001 movq %rsp,%rbp
0000000000000004 movl $0x00000041,%eax
0000000000000009 popq %rbp
000000000000000a ret
000000000000000b nopl 0x00(%rax,%rax)
_main:
0000000000000010 pushq %rbp
0000000000000011 movq %rsp,%rbp
0000000000000014 subq $0x10,%rsp
0000000000000018 movl $0x00000000,0xfc(%rbp)
000000000000001f callq 0x00000024
0000000000000024 addq $0x10,%rsp
0000000000000028 popq %rbp
0000000000000029 ret
$ llvm-rtdyld t.o -debug-only=dyld ; echo $?
Function sym: '_bar' @ 0
Function sym: '_main' @ 16
Extracting function: _bar from [0, 15]
allocated to 0x100153000
Extracting function: _main from [16, 41]
allocated to 0x100154000
Relocation at '_main' + 16 from '_bar(Word1: 0x2d000000)
Resolving relocation at '_main' + 16 (0x100154010) from '_bar (0x100153000)(pcrel, type: 2, Size: 4).
loaded '_main' at: 0x100154000
65
$
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@129388 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-12 21:20:41 +00:00
|
|
|
bool resolveRelocation(uint8_t *Address, uint8_t *Value, bool isPCRel,
|
|
|
|
unsigned Type, unsigned Size);
|
|
|
|
bool resolveX86_64Relocation(uintptr_t Address, uintptr_t Value, bool isPCRel,
|
2011-03-23 22:06:06 +00:00
|
|
|
unsigned Type, unsigned Size);
|
MCJIT lazy relocation resolution and symbol address re-assignment.
Add handling for tracking the relocations on symbols and resolving them.
Keep track of the relocations even after they are resolved so that if
the RuntimeDyld client moves the object, it can update the address and any
relocations to that object will be updated.
For our trival object file load/run test harness (llvm-rtdyld), this enables
relocations between functions located in the same object module. It should
be trivially extendable to load multiple objects with mutual references.
As a simple example, the following now works (running on x86_64 Darwin 10.6):
$ cat t.c
int bar() {
return 65;
}
int main() {
return bar();
}
$ clang t.c -fno-asynchronous-unwind-tables -o t.o -c
$ otool -vt t.o
t.o:
(__TEXT,__text) section
_bar:
0000000000000000 pushq %rbp
0000000000000001 movq %rsp,%rbp
0000000000000004 movl $0x00000041,%eax
0000000000000009 popq %rbp
000000000000000a ret
000000000000000b nopl 0x00(%rax,%rax)
_main:
0000000000000010 pushq %rbp
0000000000000011 movq %rsp,%rbp
0000000000000014 subq $0x10,%rsp
0000000000000018 movl $0x00000000,0xfc(%rbp)
000000000000001f callq 0x00000024
0000000000000024 addq $0x10,%rsp
0000000000000028 popq %rbp
0000000000000029 ret
$ llvm-rtdyld t.o -debug-only=dyld ; echo $?
Function sym: '_bar' @ 0
Function sym: '_main' @ 16
Extracting function: _bar from [0, 15]
allocated to 0x100153000
Extracting function: _main from [16, 41]
allocated to 0x100154000
Relocation at '_main' + 16 from '_bar(Word1: 0x2d000000)
Resolving relocation at '_main' + 16 (0x100154010) from '_bar (0x100153000)(pcrel, type: 2, Size: 4).
loaded '_main' at: 0x100154000
65
$
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@129388 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-12 21:20:41 +00:00
|
|
|
bool resolveARMRelocation(uintptr_t Address, uintptr_t Value, bool isPCRel,
|
2011-03-23 22:06:06 +00:00
|
|
|
unsigned Type, unsigned Size);
|
2011-03-23 19:52:00 +00:00
|
|
|
|
2011-03-21 22:15:52 +00:00
|
|
|
bool loadSegment32(const MachOObject *Obj,
|
|
|
|
const MachOObject::LoadCommandInfo *SegmentLCI,
|
|
|
|
const InMemoryStruct<macho::SymtabLoadCommand> &SymtabLC);
|
|
|
|
bool loadSegment64(const MachOObject *Obj,
|
|
|
|
const MachOObject::LoadCommandInfo *SegmentLCI,
|
|
|
|
const InMemoryStruct<macho::SymtabLoadCommand> &SymtabLC);
|
|
|
|
|
|
|
|
public:
|
2011-04-04 23:04:39 +00:00
|
|
|
RuntimeDyldImpl(RTDyldMemoryManager *mm) : MemMgr(mm), HasError(false) {}
|
2011-03-22 00:42:19 +00:00
|
|
|
|
2011-03-21 22:15:52 +00:00
|
|
|
bool loadObject(MemoryBuffer *InputBuffer);
|
|
|
|
|
2011-04-08 17:31:24 +00:00
|
|
|
void *getSymbolAddress(StringRef Name) {
|
2011-04-06 01:11:05 +00:00
|
|
|
// FIXME: Just look up as a function for now. Overly simple of course.
|
|
|
|
// Work in progress.
|
MCJIT lazy relocation resolution and symbol address re-assignment.
Add handling for tracking the relocations on symbols and resolving them.
Keep track of the relocations even after they are resolved so that if
the RuntimeDyld client moves the object, it can update the address and any
relocations to that object will be updated.
For our trival object file load/run test harness (llvm-rtdyld), this enables
relocations between functions located in the same object module. It should
be trivially extendable to load multiple objects with mutual references.
As a simple example, the following now works (running on x86_64 Darwin 10.6):
$ cat t.c
int bar() {
return 65;
}
int main() {
return bar();
}
$ clang t.c -fno-asynchronous-unwind-tables -o t.o -c
$ otool -vt t.o
t.o:
(__TEXT,__text) section
_bar:
0000000000000000 pushq %rbp
0000000000000001 movq %rsp,%rbp
0000000000000004 movl $0x00000041,%eax
0000000000000009 popq %rbp
000000000000000a ret
000000000000000b nopl 0x00(%rax,%rax)
_main:
0000000000000010 pushq %rbp
0000000000000011 movq %rsp,%rbp
0000000000000014 subq $0x10,%rsp
0000000000000018 movl $0x00000000,0xfc(%rbp)
000000000000001f callq 0x00000024
0000000000000024 addq $0x10,%rsp
0000000000000028 popq %rbp
0000000000000029 ret
$ llvm-rtdyld t.o -debug-only=dyld ; echo $?
Function sym: '_bar' @ 0
Function sym: '_main' @ 16
Extracting function: _bar from [0, 15]
allocated to 0x100153000
Extracting function: _main from [16, 41]
allocated to 0x100154000
Relocation at '_main' + 16 from '_bar(Word1: 0x2d000000)
Resolving relocation at '_main' + 16 (0x100154010) from '_bar (0x100153000)(pcrel, type: 2, Size: 4).
loaded '_main' at: 0x100154000
65
$
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@129388 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-12 21:20:41 +00:00
|
|
|
return SymbolTable.lookup(Name);
|
2011-03-21 22:15:52 +00:00
|
|
|
}
|
|
|
|
|
MCJIT lazy relocation resolution and symbol address re-assignment.
Add handling for tracking the relocations on symbols and resolving them.
Keep track of the relocations even after they are resolved so that if
the RuntimeDyld client moves the object, it can update the address and any
relocations to that object will be updated.
For our trival object file load/run test harness (llvm-rtdyld), this enables
relocations between functions located in the same object module. It should
be trivially extendable to load multiple objects with mutual references.
As a simple example, the following now works (running on x86_64 Darwin 10.6):
$ cat t.c
int bar() {
return 65;
}
int main() {
return bar();
}
$ clang t.c -fno-asynchronous-unwind-tables -o t.o -c
$ otool -vt t.o
t.o:
(__TEXT,__text) section
_bar:
0000000000000000 pushq %rbp
0000000000000001 movq %rsp,%rbp
0000000000000004 movl $0x00000041,%eax
0000000000000009 popq %rbp
000000000000000a ret
000000000000000b nopl 0x00(%rax,%rax)
_main:
0000000000000010 pushq %rbp
0000000000000011 movq %rsp,%rbp
0000000000000014 subq $0x10,%rsp
0000000000000018 movl $0x00000000,0xfc(%rbp)
000000000000001f callq 0x00000024
0000000000000024 addq $0x10,%rsp
0000000000000028 popq %rbp
0000000000000029 ret
$ llvm-rtdyld t.o -debug-only=dyld ; echo $?
Function sym: '_bar' @ 0
Function sym: '_main' @ 16
Extracting function: _bar from [0, 15]
allocated to 0x100153000
Extracting function: _main from [16, 41]
allocated to 0x100154000
Relocation at '_main' + 16 from '_bar(Word1: 0x2d000000)
Resolving relocation at '_main' + 16 (0x100154010) from '_bar (0x100153000)(pcrel, type: 2, Size: 4).
loaded '_main' at: 0x100154000
65
$
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@129388 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-12 21:20:41 +00:00
|
|
|
void resolveRelocations();
|
|
|
|
|
|
|
|
void reassignSymbolAddress(StringRef Name, uint8_t *Addr);
|
|
|
|
|
2011-03-21 22:15:52 +00:00
|
|
|
// Is the linker in an error state?
|
|
|
|
bool hasError() { return HasError; }
|
|
|
|
|
|
|
|
// Mark the error condition as handled and continue.
|
|
|
|
void clearError() { HasError = false; }
|
|
|
|
|
|
|
|
// Get the error message.
|
|
|
|
StringRef getErrorString() { return ErrorStr; }
|
|
|
|
};
|
|
|
|
|
2011-04-06 01:11:05 +00:00
|
|
|
void RuntimeDyldImpl::extractFunction(StringRef Name, uint8_t *StartAddress,
|
2011-04-06 22:13:52 +00:00
|
|
|
uint8_t *EndAddress) {
|
2011-04-06 01:11:05 +00:00
|
|
|
// Allocate memory for the function via the memory manager.
|
|
|
|
uintptr_t Size = EndAddress - StartAddress + 1;
|
|
|
|
uint8_t *Mem = MemMgr->startFunctionBody(Name.data(), Size);
|
|
|
|
assert(Size >= (uint64_t)(EndAddress - StartAddress + 1) &&
|
|
|
|
"Memory manager failed to allocate enough memory!");
|
|
|
|
// Copy the function payload into the memory block.
|
|
|
|
memcpy(Mem, StartAddress, EndAddress - StartAddress + 1);
|
|
|
|
MemMgr->endFunctionBody(Name.data(), Mem, Mem + Size);
|
|
|
|
// Remember where we put it.
|
|
|
|
Functions[Name] = sys::MemoryBlock(Mem, Size);
|
MCJIT lazy relocation resolution and symbol address re-assignment.
Add handling for tracking the relocations on symbols and resolving them.
Keep track of the relocations even after they are resolved so that if
the RuntimeDyld client moves the object, it can update the address and any
relocations to that object will be updated.
For our trival object file load/run test harness (llvm-rtdyld), this enables
relocations between functions located in the same object module. It should
be trivially extendable to load multiple objects with mutual references.
As a simple example, the following now works (running on x86_64 Darwin 10.6):
$ cat t.c
int bar() {
return 65;
}
int main() {
return bar();
}
$ clang t.c -fno-asynchronous-unwind-tables -o t.o -c
$ otool -vt t.o
t.o:
(__TEXT,__text) section
_bar:
0000000000000000 pushq %rbp
0000000000000001 movq %rsp,%rbp
0000000000000004 movl $0x00000041,%eax
0000000000000009 popq %rbp
000000000000000a ret
000000000000000b nopl 0x00(%rax,%rax)
_main:
0000000000000010 pushq %rbp
0000000000000011 movq %rsp,%rbp
0000000000000014 subq $0x10,%rsp
0000000000000018 movl $0x00000000,0xfc(%rbp)
000000000000001f callq 0x00000024
0000000000000024 addq $0x10,%rsp
0000000000000028 popq %rbp
0000000000000029 ret
$ llvm-rtdyld t.o -debug-only=dyld ; echo $?
Function sym: '_bar' @ 0
Function sym: '_main' @ 16
Extracting function: _bar from [0, 15]
allocated to 0x100153000
Extracting function: _main from [16, 41]
allocated to 0x100154000
Relocation at '_main' + 16 from '_bar(Word1: 0x2d000000)
Resolving relocation at '_main' + 16 (0x100154010) from '_bar (0x100153000)(pcrel, type: 2, Size: 4).
loaded '_main' at: 0x100154000
65
$
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@129388 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-12 21:20:41 +00:00
|
|
|
// Default the assigned address for this symbol to wherever this
|
|
|
|
// allocated it.
|
|
|
|
SymbolTable[Name] = Mem;
|
2011-04-06 01:11:05 +00:00
|
|
|
DEBUG(dbgs() << " allocated to " << Mem << "\n");
|
|
|
|
}
|
|
|
|
|
2011-03-23 19:52:00 +00:00
|
|
|
bool RuntimeDyldImpl::
|
MCJIT lazy relocation resolution and symbol address re-assignment.
Add handling for tracking the relocations on symbols and resolving them.
Keep track of the relocations even after they are resolved so that if
the RuntimeDyld client moves the object, it can update the address and any
relocations to that object will be updated.
For our trival object file load/run test harness (llvm-rtdyld), this enables
relocations between functions located in the same object module. It should
be trivially extendable to load multiple objects with mutual references.
As a simple example, the following now works (running on x86_64 Darwin 10.6):
$ cat t.c
int bar() {
return 65;
}
int main() {
return bar();
}
$ clang t.c -fno-asynchronous-unwind-tables -o t.o -c
$ otool -vt t.o
t.o:
(__TEXT,__text) section
_bar:
0000000000000000 pushq %rbp
0000000000000001 movq %rsp,%rbp
0000000000000004 movl $0x00000041,%eax
0000000000000009 popq %rbp
000000000000000a ret
000000000000000b nopl 0x00(%rax,%rax)
_main:
0000000000000010 pushq %rbp
0000000000000011 movq %rsp,%rbp
0000000000000014 subq $0x10,%rsp
0000000000000018 movl $0x00000000,0xfc(%rbp)
000000000000001f callq 0x00000024
0000000000000024 addq $0x10,%rsp
0000000000000028 popq %rbp
0000000000000029 ret
$ llvm-rtdyld t.o -debug-only=dyld ; echo $?
Function sym: '_bar' @ 0
Function sym: '_main' @ 16
Extracting function: _bar from [0, 15]
allocated to 0x100153000
Extracting function: _main from [16, 41]
allocated to 0x100154000
Relocation at '_main' + 16 from '_bar(Word1: 0x2d000000)
Resolving relocation at '_main' + 16 (0x100154010) from '_bar (0x100153000)(pcrel, type: 2, Size: 4).
loaded '_main' at: 0x100154000
65
$
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@129388 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-12 21:20:41 +00:00
|
|
|
resolveRelocation(uint8_t *Address, uint8_t *Value, bool isPCRel,
|
|
|
|
unsigned Type, unsigned Size) {
|
|
|
|
// This just dispatches to the proper target specific routine.
|
2011-03-23 22:06:06 +00:00
|
|
|
switch (CPUType) {
|
|
|
|
default: assert(0 && "Unsupported CPU type!");
|
|
|
|
case mach::CTM_x86_64:
|
MCJIT lazy relocation resolution and symbol address re-assignment.
Add handling for tracking the relocations on symbols and resolving them.
Keep track of the relocations even after they are resolved so that if
the RuntimeDyld client moves the object, it can update the address and any
relocations to that object will be updated.
For our trival object file load/run test harness (llvm-rtdyld), this enables
relocations between functions located in the same object module. It should
be trivially extendable to load multiple objects with mutual references.
As a simple example, the following now works (running on x86_64 Darwin 10.6):
$ cat t.c
int bar() {
return 65;
}
int main() {
return bar();
}
$ clang t.c -fno-asynchronous-unwind-tables -o t.o -c
$ otool -vt t.o
t.o:
(__TEXT,__text) section
_bar:
0000000000000000 pushq %rbp
0000000000000001 movq %rsp,%rbp
0000000000000004 movl $0x00000041,%eax
0000000000000009 popq %rbp
000000000000000a ret
000000000000000b nopl 0x00(%rax,%rax)
_main:
0000000000000010 pushq %rbp
0000000000000011 movq %rsp,%rbp
0000000000000014 subq $0x10,%rsp
0000000000000018 movl $0x00000000,0xfc(%rbp)
000000000000001f callq 0x00000024
0000000000000024 addq $0x10,%rsp
0000000000000028 popq %rbp
0000000000000029 ret
$ llvm-rtdyld t.o -debug-only=dyld ; echo $?
Function sym: '_bar' @ 0
Function sym: '_main' @ 16
Extracting function: _bar from [0, 15]
allocated to 0x100153000
Extracting function: _main from [16, 41]
allocated to 0x100154000
Relocation at '_main' + 16 from '_bar(Word1: 0x2d000000)
Resolving relocation at '_main' + 16 (0x100154010) from '_bar (0x100153000)(pcrel, type: 2, Size: 4).
loaded '_main' at: 0x100154000
65
$
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@129388 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-12 21:20:41 +00:00
|
|
|
return resolveX86_64Relocation((uintptr_t)Address, (uintptr_t)Value,
|
|
|
|
isPCRel, Type, Size);
|
2011-03-23 22:06:06 +00:00
|
|
|
case mach::CTM_ARM:
|
MCJIT lazy relocation resolution and symbol address re-assignment.
Add handling for tracking the relocations on symbols and resolving them.
Keep track of the relocations even after they are resolved so that if
the RuntimeDyld client moves the object, it can update the address and any
relocations to that object will be updated.
For our trival object file load/run test harness (llvm-rtdyld), this enables
relocations between functions located in the same object module. It should
be trivially extendable to load multiple objects with mutual references.
As a simple example, the following now works (running on x86_64 Darwin 10.6):
$ cat t.c
int bar() {
return 65;
}
int main() {
return bar();
}
$ clang t.c -fno-asynchronous-unwind-tables -o t.o -c
$ otool -vt t.o
t.o:
(__TEXT,__text) section
_bar:
0000000000000000 pushq %rbp
0000000000000001 movq %rsp,%rbp
0000000000000004 movl $0x00000041,%eax
0000000000000009 popq %rbp
000000000000000a ret
000000000000000b nopl 0x00(%rax,%rax)
_main:
0000000000000010 pushq %rbp
0000000000000011 movq %rsp,%rbp
0000000000000014 subq $0x10,%rsp
0000000000000018 movl $0x00000000,0xfc(%rbp)
000000000000001f callq 0x00000024
0000000000000024 addq $0x10,%rsp
0000000000000028 popq %rbp
0000000000000029 ret
$ llvm-rtdyld t.o -debug-only=dyld ; echo $?
Function sym: '_bar' @ 0
Function sym: '_main' @ 16
Extracting function: _bar from [0, 15]
allocated to 0x100153000
Extracting function: _main from [16, 41]
allocated to 0x100154000
Relocation at '_main' + 16 from '_bar(Word1: 0x2d000000)
Resolving relocation at '_main' + 16 (0x100154010) from '_bar (0x100153000)(pcrel, type: 2, Size: 4).
loaded '_main' at: 0x100154000
65
$
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@129388 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-12 21:20:41 +00:00
|
|
|
return resolveARMRelocation((uintptr_t)Address, (uintptr_t)Value,
|
|
|
|
isPCRel, Type, Size);
|
2011-03-23 22:06:06 +00:00
|
|
|
}
|
|
|
|
llvm_unreachable("");
|
|
|
|
}
|
|
|
|
|
MCJIT lazy relocation resolution and symbol address re-assignment.
Add handling for tracking the relocations on symbols and resolving them.
Keep track of the relocations even after they are resolved so that if
the RuntimeDyld client moves the object, it can update the address and any
relocations to that object will be updated.
For our trival object file load/run test harness (llvm-rtdyld), this enables
relocations between functions located in the same object module. It should
be trivially extendable to load multiple objects with mutual references.
As a simple example, the following now works (running on x86_64 Darwin 10.6):
$ cat t.c
int bar() {
return 65;
}
int main() {
return bar();
}
$ clang t.c -fno-asynchronous-unwind-tables -o t.o -c
$ otool -vt t.o
t.o:
(__TEXT,__text) section
_bar:
0000000000000000 pushq %rbp
0000000000000001 movq %rsp,%rbp
0000000000000004 movl $0x00000041,%eax
0000000000000009 popq %rbp
000000000000000a ret
000000000000000b nopl 0x00(%rax,%rax)
_main:
0000000000000010 pushq %rbp
0000000000000011 movq %rsp,%rbp
0000000000000014 subq $0x10,%rsp
0000000000000018 movl $0x00000000,0xfc(%rbp)
000000000000001f callq 0x00000024
0000000000000024 addq $0x10,%rsp
0000000000000028 popq %rbp
0000000000000029 ret
$ llvm-rtdyld t.o -debug-only=dyld ; echo $?
Function sym: '_bar' @ 0
Function sym: '_main' @ 16
Extracting function: _bar from [0, 15]
allocated to 0x100153000
Extracting function: _main from [16, 41]
allocated to 0x100154000
Relocation at '_main' + 16 from '_bar(Word1: 0x2d000000)
Resolving relocation at '_main' + 16 (0x100154010) from '_bar (0x100153000)(pcrel, type: 2, Size: 4).
loaded '_main' at: 0x100154000
65
$
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@129388 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-12 21:20:41 +00:00
|
|
|
bool RuntimeDyldImpl::
|
|
|
|
resolveX86_64Relocation(uintptr_t Address, uintptr_t Value,
|
|
|
|
bool isPCRel, unsigned Type,
|
|
|
|
unsigned Size) {
|
2011-03-23 19:52:00 +00:00
|
|
|
// If the relocation is PC-relative, the value to be encoded is the
|
|
|
|
// pointer difference.
|
|
|
|
if (isPCRel)
|
|
|
|
// FIXME: It seems this value needs to be adjusted by 4 for an effective PC
|
|
|
|
// address. Is that expected? Only for branches, perhaps?
|
|
|
|
Value -= Address + 4;
|
|
|
|
|
|
|
|
switch(Type) {
|
|
|
|
default:
|
|
|
|
llvm_unreachable("Invalid relocation type!");
|
|
|
|
case macho::RIT_X86_64_Unsigned:
|
|
|
|
case macho::RIT_X86_64_Branch: {
|
|
|
|
// Mask in the target value a byte at a time (we don't have an alignment
|
|
|
|
// guarantee for the target address, so this is safest).
|
|
|
|
uint8_t *p = (uint8_t*)Address;
|
2011-03-23 22:06:06 +00:00
|
|
|
for (unsigned i = 0; i < Size; ++i) {
|
2011-03-23 19:52:00 +00:00
|
|
|
*p++ = (uint8_t)Value;
|
|
|
|
Value >>= 8;
|
|
|
|
}
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
case macho::RIT_X86_64_Signed:
|
|
|
|
case macho::RIT_X86_64_GOTLoad:
|
|
|
|
case macho::RIT_X86_64_GOT:
|
|
|
|
case macho::RIT_X86_64_Subtractor:
|
|
|
|
case macho::RIT_X86_64_Signed1:
|
|
|
|
case macho::RIT_X86_64_Signed2:
|
|
|
|
case macho::RIT_X86_64_Signed4:
|
|
|
|
case macho::RIT_X86_64_TLV:
|
|
|
|
return Error("Relocation type not implemented yet!");
|
|
|
|
}
|
|
|
|
return false;
|
|
|
|
}
|
2011-03-21 22:15:52 +00:00
|
|
|
|
MCJIT lazy relocation resolution and symbol address re-assignment.
Add handling for tracking the relocations on symbols and resolving them.
Keep track of the relocations even after they are resolved so that if
the RuntimeDyld client moves the object, it can update the address and any
relocations to that object will be updated.
For our trival object file load/run test harness (llvm-rtdyld), this enables
relocations between functions located in the same object module. It should
be trivially extendable to load multiple objects with mutual references.
As a simple example, the following now works (running on x86_64 Darwin 10.6):
$ cat t.c
int bar() {
return 65;
}
int main() {
return bar();
}
$ clang t.c -fno-asynchronous-unwind-tables -o t.o -c
$ otool -vt t.o
t.o:
(__TEXT,__text) section
_bar:
0000000000000000 pushq %rbp
0000000000000001 movq %rsp,%rbp
0000000000000004 movl $0x00000041,%eax
0000000000000009 popq %rbp
000000000000000a ret
000000000000000b nopl 0x00(%rax,%rax)
_main:
0000000000000010 pushq %rbp
0000000000000011 movq %rsp,%rbp
0000000000000014 subq $0x10,%rsp
0000000000000018 movl $0x00000000,0xfc(%rbp)
000000000000001f callq 0x00000024
0000000000000024 addq $0x10,%rsp
0000000000000028 popq %rbp
0000000000000029 ret
$ llvm-rtdyld t.o -debug-only=dyld ; echo $?
Function sym: '_bar' @ 0
Function sym: '_main' @ 16
Extracting function: _bar from [0, 15]
allocated to 0x100153000
Extracting function: _main from [16, 41]
allocated to 0x100154000
Relocation at '_main' + 16 from '_bar(Word1: 0x2d000000)
Resolving relocation at '_main' + 16 (0x100154010) from '_bar (0x100153000)(pcrel, type: 2, Size: 4).
loaded '_main' at: 0x100154000
65
$
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@129388 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-12 21:20:41 +00:00
|
|
|
bool RuntimeDyldImpl::resolveARMRelocation(uintptr_t Address, uintptr_t Value,
|
2011-03-23 22:06:06 +00:00
|
|
|
bool isPCRel, unsigned Type,
|
|
|
|
unsigned Size) {
|
|
|
|
// If the relocation is PC-relative, the value to be encoded is the
|
|
|
|
// pointer difference.
|
|
|
|
if (isPCRel) {
|
|
|
|
Value -= Address;
|
|
|
|
// ARM PCRel relocations have an effective-PC offset of two instructions
|
|
|
|
// (four bytes in Thumb mode, 8 bytes in ARM mode).
|
|
|
|
// FIXME: For now, assume ARM mode.
|
|
|
|
Value -= 8;
|
|
|
|
}
|
|
|
|
|
|
|
|
switch(Type) {
|
|
|
|
default:
|
MCJIT lazy relocation resolution and symbol address re-assignment.
Add handling for tracking the relocations on symbols and resolving them.
Keep track of the relocations even after they are resolved so that if
the RuntimeDyld client moves the object, it can update the address and any
relocations to that object will be updated.
For our trival object file load/run test harness (llvm-rtdyld), this enables
relocations between functions located in the same object module. It should
be trivially extendable to load multiple objects with mutual references.
As a simple example, the following now works (running on x86_64 Darwin 10.6):
$ cat t.c
int bar() {
return 65;
}
int main() {
return bar();
}
$ clang t.c -fno-asynchronous-unwind-tables -o t.o -c
$ otool -vt t.o
t.o:
(__TEXT,__text) section
_bar:
0000000000000000 pushq %rbp
0000000000000001 movq %rsp,%rbp
0000000000000004 movl $0x00000041,%eax
0000000000000009 popq %rbp
000000000000000a ret
000000000000000b nopl 0x00(%rax,%rax)
_main:
0000000000000010 pushq %rbp
0000000000000011 movq %rsp,%rbp
0000000000000014 subq $0x10,%rsp
0000000000000018 movl $0x00000000,0xfc(%rbp)
000000000000001f callq 0x00000024
0000000000000024 addq $0x10,%rsp
0000000000000028 popq %rbp
0000000000000029 ret
$ llvm-rtdyld t.o -debug-only=dyld ; echo $?
Function sym: '_bar' @ 0
Function sym: '_main' @ 16
Extracting function: _bar from [0, 15]
allocated to 0x100153000
Extracting function: _main from [16, 41]
allocated to 0x100154000
Relocation at '_main' + 16 from '_bar(Word1: 0x2d000000)
Resolving relocation at '_main' + 16 (0x100154010) from '_bar (0x100153000)(pcrel, type: 2, Size: 4).
loaded '_main' at: 0x100154000
65
$
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@129388 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-12 21:20:41 +00:00
|
|
|
llvm_unreachable("Invalid relocation type!");
|
2011-03-23 22:06:06 +00:00
|
|
|
case macho::RIT_Vanilla: {
|
|
|
|
llvm_unreachable("Invalid relocation type!");
|
|
|
|
// Mask in the target value a byte at a time (we don't have an alignment
|
|
|
|
// guarantee for the target address, so this is safest).
|
|
|
|
uint8_t *p = (uint8_t*)Address;
|
|
|
|
for (unsigned i = 0; i < Size; ++i) {
|
|
|
|
*p++ = (uint8_t)Value;
|
|
|
|
Value >>= 8;
|
|
|
|
}
|
2011-03-23 23:35:17 +00:00
|
|
|
break;
|
2011-03-23 22:06:06 +00:00
|
|
|
}
|
2011-03-23 23:35:17 +00:00
|
|
|
case macho::RIT_ARM_Branch24Bit: {
|
|
|
|
// Mask the value into the target address. We know instructions are
|
|
|
|
// 32-bit aligned, so we can do it all at once.
|
|
|
|
uint32_t *p = (uint32_t*)Address;
|
|
|
|
// The low two bits of the value are not encoded.
|
|
|
|
Value >>= 2;
|
|
|
|
// Mask the value to 24 bits.
|
|
|
|
Value &= 0xffffff;
|
|
|
|
// FIXME: If the destination is a Thumb function (and the instruction
|
|
|
|
// is a non-predicated BL instruction), we need to change it to a BLX
|
|
|
|
// instruction instead.
|
|
|
|
|
|
|
|
// Insert the value into the instruction.
|
|
|
|
*p = (*p & ~0xffffff) | Value;
|
|
|
|
break;
|
|
|
|
}
|
2011-03-23 22:06:06 +00:00
|
|
|
case macho::RIT_ARM_ThumbBranch22Bit:
|
|
|
|
case macho::RIT_ARM_ThumbBranch32Bit:
|
|
|
|
case macho::RIT_ARM_Half:
|
|
|
|
case macho::RIT_ARM_HalfDifference:
|
MCJIT lazy relocation resolution and symbol address re-assignment.
Add handling for tracking the relocations on symbols and resolving them.
Keep track of the relocations even after they are resolved so that if
the RuntimeDyld client moves the object, it can update the address and any
relocations to that object will be updated.
For our trival object file load/run test harness (llvm-rtdyld), this enables
relocations between functions located in the same object module. It should
be trivially extendable to load multiple objects with mutual references.
As a simple example, the following now works (running on x86_64 Darwin 10.6):
$ cat t.c
int bar() {
return 65;
}
int main() {
return bar();
}
$ clang t.c -fno-asynchronous-unwind-tables -o t.o -c
$ otool -vt t.o
t.o:
(__TEXT,__text) section
_bar:
0000000000000000 pushq %rbp
0000000000000001 movq %rsp,%rbp
0000000000000004 movl $0x00000041,%eax
0000000000000009 popq %rbp
000000000000000a ret
000000000000000b nopl 0x00(%rax,%rax)
_main:
0000000000000010 pushq %rbp
0000000000000011 movq %rsp,%rbp
0000000000000014 subq $0x10,%rsp
0000000000000018 movl $0x00000000,0xfc(%rbp)
000000000000001f callq 0x00000024
0000000000000024 addq $0x10,%rsp
0000000000000028 popq %rbp
0000000000000029 ret
$ llvm-rtdyld t.o -debug-only=dyld ; echo $?
Function sym: '_bar' @ 0
Function sym: '_main' @ 16
Extracting function: _bar from [0, 15]
allocated to 0x100153000
Extracting function: _main from [16, 41]
allocated to 0x100154000
Relocation at '_main' + 16 from '_bar(Word1: 0x2d000000)
Resolving relocation at '_main' + 16 (0x100154010) from '_bar (0x100153000)(pcrel, type: 2, Size: 4).
loaded '_main' at: 0x100154000
65
$
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@129388 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-12 21:20:41 +00:00
|
|
|
case macho::RIT_Pair:
|
|
|
|
case macho::RIT_Difference:
|
|
|
|
case macho::RIT_ARM_LocalDifference:
|
|
|
|
case macho::RIT_ARM_PreboundLazyPointer:
|
2011-03-23 22:06:06 +00:00
|
|
|
return Error("Relocation type not implemented yet!");
|
|
|
|
}
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
2011-03-21 22:15:52 +00:00
|
|
|
bool RuntimeDyldImpl::
|
|
|
|
loadSegment32(const MachOObject *Obj,
|
|
|
|
const MachOObject::LoadCommandInfo *SegmentLCI,
|
|
|
|
const InMemoryStruct<macho::SymtabLoadCommand> &SymtabLC) {
|
MCJIT lazy relocation resolution and symbol address re-assignment.
Add handling for tracking the relocations on symbols and resolving them.
Keep track of the relocations even after they are resolved so that if
the RuntimeDyld client moves the object, it can update the address and any
relocations to that object will be updated.
For our trival object file load/run test harness (llvm-rtdyld), this enables
relocations between functions located in the same object module. It should
be trivially extendable to load multiple objects with mutual references.
As a simple example, the following now works (running on x86_64 Darwin 10.6):
$ cat t.c
int bar() {
return 65;
}
int main() {
return bar();
}
$ clang t.c -fno-asynchronous-unwind-tables -o t.o -c
$ otool -vt t.o
t.o:
(__TEXT,__text) section
_bar:
0000000000000000 pushq %rbp
0000000000000001 movq %rsp,%rbp
0000000000000004 movl $0x00000041,%eax
0000000000000009 popq %rbp
000000000000000a ret
000000000000000b nopl 0x00(%rax,%rax)
_main:
0000000000000010 pushq %rbp
0000000000000011 movq %rsp,%rbp
0000000000000014 subq $0x10,%rsp
0000000000000018 movl $0x00000000,0xfc(%rbp)
000000000000001f callq 0x00000024
0000000000000024 addq $0x10,%rsp
0000000000000028 popq %rbp
0000000000000029 ret
$ llvm-rtdyld t.o -debug-only=dyld ; echo $?
Function sym: '_bar' @ 0
Function sym: '_main' @ 16
Extracting function: _bar from [0, 15]
allocated to 0x100153000
Extracting function: _main from [16, 41]
allocated to 0x100154000
Relocation at '_main' + 16 from '_bar(Word1: 0x2d000000)
Resolving relocation at '_main' + 16 (0x100154010) from '_bar (0x100153000)(pcrel, type: 2, Size: 4).
loaded '_main' at: 0x100154000
65
$
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@129388 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-12 21:20:41 +00:00
|
|
|
InMemoryStruct<macho::SegmentLoadCommand> SegmentLC;
|
|
|
|
Obj->ReadSegmentLoadCommand(*SegmentLCI, SegmentLC);
|
|
|
|
if (!SegmentLC)
|
2011-03-21 22:15:52 +00:00
|
|
|
return Error("unable to load segment load command");
|
|
|
|
|
MCJIT lazy relocation resolution and symbol address re-assignment.
Add handling for tracking the relocations on symbols and resolving them.
Keep track of the relocations even after they are resolved so that if
the RuntimeDyld client moves the object, it can update the address and any
relocations to that object will be updated.
For our trival object file load/run test harness (llvm-rtdyld), this enables
relocations between functions located in the same object module. It should
be trivially extendable to load multiple objects with mutual references.
As a simple example, the following now works (running on x86_64 Darwin 10.6):
$ cat t.c
int bar() {
return 65;
}
int main() {
return bar();
}
$ clang t.c -fno-asynchronous-unwind-tables -o t.o -c
$ otool -vt t.o
t.o:
(__TEXT,__text) section
_bar:
0000000000000000 pushq %rbp
0000000000000001 movq %rsp,%rbp
0000000000000004 movl $0x00000041,%eax
0000000000000009 popq %rbp
000000000000000a ret
000000000000000b nopl 0x00(%rax,%rax)
_main:
0000000000000010 pushq %rbp
0000000000000011 movq %rsp,%rbp
0000000000000014 subq $0x10,%rsp
0000000000000018 movl $0x00000000,0xfc(%rbp)
000000000000001f callq 0x00000024
0000000000000024 addq $0x10,%rsp
0000000000000028 popq %rbp
0000000000000029 ret
$ llvm-rtdyld t.o -debug-only=dyld ; echo $?
Function sym: '_bar' @ 0
Function sym: '_main' @ 16
Extracting function: _bar from [0, 15]
allocated to 0x100153000
Extracting function: _main from [16, 41]
allocated to 0x100154000
Relocation at '_main' + 16 from '_bar(Word1: 0x2d000000)
Resolving relocation at '_main' + 16 (0x100154010) from '_bar (0x100153000)(pcrel, type: 2, Size: 4).
loaded '_main' at: 0x100154000
65
$
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@129388 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-12 21:20:41 +00:00
|
|
|
for (unsigned SectNum = 0; SectNum != SegmentLC->NumSections; ++SectNum) {
|
2011-03-21 22:15:52 +00:00
|
|
|
InMemoryStruct<macho::Section> Sect;
|
2011-04-08 17:31:24 +00:00
|
|
|
Obj->ReadSection(*SegmentLCI, SectNum, Sect);
|
|
|
|
if (!Sect)
|
|
|
|
return Error("unable to load section: '" + Twine(SectNum) + "'");
|
2011-03-21 22:15:52 +00:00
|
|
|
|
|
|
|
// FIXME: Improve check.
|
2011-04-08 17:31:24 +00:00
|
|
|
if (Sect->Flags != 0x80000400)
|
|
|
|
return Error("unsupported section type!");
|
2011-03-21 22:15:52 +00:00
|
|
|
|
2011-04-08 17:31:24 +00:00
|
|
|
// Address and names of symbols in the section.
|
|
|
|
typedef std::pair<uint64_t, StringRef> SymbolEntry;
|
MCJIT lazy relocation resolution and symbol address re-assignment.
Add handling for tracking the relocations on symbols and resolving them.
Keep track of the relocations even after they are resolved so that if
the RuntimeDyld client moves the object, it can update the address and any
relocations to that object will be updated.
For our trival object file load/run test harness (llvm-rtdyld), this enables
relocations between functions located in the same object module. It should
be trivially extendable to load multiple objects with mutual references.
As a simple example, the following now works (running on x86_64 Darwin 10.6):
$ cat t.c
int bar() {
return 65;
}
int main() {
return bar();
}
$ clang t.c -fno-asynchronous-unwind-tables -o t.o -c
$ otool -vt t.o
t.o:
(__TEXT,__text) section
_bar:
0000000000000000 pushq %rbp
0000000000000001 movq %rsp,%rbp
0000000000000004 movl $0x00000041,%eax
0000000000000009 popq %rbp
000000000000000a ret
000000000000000b nopl 0x00(%rax,%rax)
_main:
0000000000000010 pushq %rbp
0000000000000011 movq %rsp,%rbp
0000000000000014 subq $0x10,%rsp
0000000000000018 movl $0x00000000,0xfc(%rbp)
000000000000001f callq 0x00000024
0000000000000024 addq $0x10,%rsp
0000000000000028 popq %rbp
0000000000000029 ret
$ llvm-rtdyld t.o -debug-only=dyld ; echo $?
Function sym: '_bar' @ 0
Function sym: '_main' @ 16
Extracting function: _bar from [0, 15]
allocated to 0x100153000
Extracting function: _main from [16, 41]
allocated to 0x100154000
Relocation at '_main' + 16 from '_bar(Word1: 0x2d000000)
Resolving relocation at '_main' + 16 (0x100154010) from '_bar (0x100153000)(pcrel, type: 2, Size: 4).
loaded '_main' at: 0x100154000
65
$
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@129388 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-12 21:20:41 +00:00
|
|
|
SmallVector<SymbolEntry, 64> Symbols;
|
|
|
|
// Index of all the names, in this section or not. Used when we're
|
|
|
|
// dealing with relocation entries.
|
|
|
|
SmallVector<StringRef, 64> SymbolNames;
|
2011-04-08 17:31:24 +00:00
|
|
|
for (unsigned i = 0; i != SymtabLC->NumSymbolTableEntries; ++i) {
|
|
|
|
InMemoryStruct<macho::SymbolTableEntry> STE;
|
|
|
|
Obj->ReadSymbolTableEntry(SymtabLC->SymbolTableOffset, i, STE);
|
|
|
|
if (!STE)
|
|
|
|
return Error("unable to read symbol: '" + Twine(i) + "'");
|
MCJIT lazy relocation resolution and symbol address re-assignment.
Add handling for tracking the relocations on symbols and resolving them.
Keep track of the relocations even after they are resolved so that if
the RuntimeDyld client moves the object, it can update the address and any
relocations to that object will be updated.
For our trival object file load/run test harness (llvm-rtdyld), this enables
relocations between functions located in the same object module. It should
be trivially extendable to load multiple objects with mutual references.
As a simple example, the following now works (running on x86_64 Darwin 10.6):
$ cat t.c
int bar() {
return 65;
}
int main() {
return bar();
}
$ clang t.c -fno-asynchronous-unwind-tables -o t.o -c
$ otool -vt t.o
t.o:
(__TEXT,__text) section
_bar:
0000000000000000 pushq %rbp
0000000000000001 movq %rsp,%rbp
0000000000000004 movl $0x00000041,%eax
0000000000000009 popq %rbp
000000000000000a ret
000000000000000b nopl 0x00(%rax,%rax)
_main:
0000000000000010 pushq %rbp
0000000000000011 movq %rsp,%rbp
0000000000000014 subq $0x10,%rsp
0000000000000018 movl $0x00000000,0xfc(%rbp)
000000000000001f callq 0x00000024
0000000000000024 addq $0x10,%rsp
0000000000000028 popq %rbp
0000000000000029 ret
$ llvm-rtdyld t.o -debug-only=dyld ; echo $?
Function sym: '_bar' @ 0
Function sym: '_main' @ 16
Extracting function: _bar from [0, 15]
allocated to 0x100153000
Extracting function: _main from [16, 41]
allocated to 0x100154000
Relocation at '_main' + 16 from '_bar(Word1: 0x2d000000)
Resolving relocation at '_main' + 16 (0x100154010) from '_bar (0x100153000)(pcrel, type: 2, Size: 4).
loaded '_main' at: 0x100154000
65
$
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@129388 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-12 21:20:41 +00:00
|
|
|
if (STE->SectionIndex > SegmentLC->NumSections)
|
2011-04-09 10:10:35 +00:00
|
|
|
return Error("invalid section index for symbol: '" + Twine(i) + "'");
|
MCJIT lazy relocation resolution and symbol address re-assignment.
Add handling for tracking the relocations on symbols and resolving them.
Keep track of the relocations even after they are resolved so that if
the RuntimeDyld client moves the object, it can update the address and any
relocations to that object will be updated.
For our trival object file load/run test harness (llvm-rtdyld), this enables
relocations between functions located in the same object module. It should
be trivially extendable to load multiple objects with mutual references.
As a simple example, the following now works (running on x86_64 Darwin 10.6):
$ cat t.c
int bar() {
return 65;
}
int main() {
return bar();
}
$ clang t.c -fno-asynchronous-unwind-tables -o t.o -c
$ otool -vt t.o
t.o:
(__TEXT,__text) section
_bar:
0000000000000000 pushq %rbp
0000000000000001 movq %rsp,%rbp
0000000000000004 movl $0x00000041,%eax
0000000000000009 popq %rbp
000000000000000a ret
000000000000000b nopl 0x00(%rax,%rax)
_main:
0000000000000010 pushq %rbp
0000000000000011 movq %rsp,%rbp
0000000000000014 subq $0x10,%rsp
0000000000000018 movl $0x00000000,0xfc(%rbp)
000000000000001f callq 0x00000024
0000000000000024 addq $0x10,%rsp
0000000000000028 popq %rbp
0000000000000029 ret
$ llvm-rtdyld t.o -debug-only=dyld ; echo $?
Function sym: '_bar' @ 0
Function sym: '_main' @ 16
Extracting function: _bar from [0, 15]
allocated to 0x100153000
Extracting function: _main from [16, 41]
allocated to 0x100154000
Relocation at '_main' + 16 from '_bar(Word1: 0x2d000000)
Resolving relocation at '_main' + 16 (0x100154010) from '_bar (0x100153000)(pcrel, type: 2, Size: 4).
loaded '_main' at: 0x100154000
65
$
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@129388 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-12 21:20:41 +00:00
|
|
|
// Get the symbol name.
|
|
|
|
StringRef Name = Obj->getStringAtIndex(STE->StringIndex);
|
|
|
|
SymbolNames.push_back(Name);
|
2011-03-21 22:15:52 +00:00
|
|
|
|
2011-04-08 17:31:24 +00:00
|
|
|
// Just skip symbols not defined in this section.
|
2011-04-08 21:11:20 +00:00
|
|
|
if ((unsigned)STE->SectionIndex - 1 != SectNum)
|
2011-04-08 17:31:24 +00:00
|
|
|
continue;
|
2011-03-21 22:15:52 +00:00
|
|
|
|
2011-04-08 17:31:24 +00:00
|
|
|
// FIXME: Check the symbol type and flags.
|
|
|
|
if (STE->Type != 0xF) // external, defined in this section.
|
|
|
|
return Error("unexpected symbol type!");
|
MCJIT lazy relocation resolution and symbol address re-assignment.
Add handling for tracking the relocations on symbols and resolving them.
Keep track of the relocations even after they are resolved so that if
the RuntimeDyld client moves the object, it can update the address and any
relocations to that object will be updated.
For our trival object file load/run test harness (llvm-rtdyld), this enables
relocations between functions located in the same object module. It should
be trivially extendable to load multiple objects with mutual references.
As a simple example, the following now works (running on x86_64 Darwin 10.6):
$ cat t.c
int bar() {
return 65;
}
int main() {
return bar();
}
$ clang t.c -fno-asynchronous-unwind-tables -o t.o -c
$ otool -vt t.o
t.o:
(__TEXT,__text) section
_bar:
0000000000000000 pushq %rbp
0000000000000001 movq %rsp,%rbp
0000000000000004 movl $0x00000041,%eax
0000000000000009 popq %rbp
000000000000000a ret
000000000000000b nopl 0x00(%rax,%rax)
_main:
0000000000000010 pushq %rbp
0000000000000011 movq %rsp,%rbp
0000000000000014 subq $0x10,%rsp
0000000000000018 movl $0x00000000,0xfc(%rbp)
000000000000001f callq 0x00000024
0000000000000024 addq $0x10,%rsp
0000000000000028 popq %rbp
0000000000000029 ret
$ llvm-rtdyld t.o -debug-only=dyld ; echo $?
Function sym: '_bar' @ 0
Function sym: '_main' @ 16
Extracting function: _bar from [0, 15]
allocated to 0x100153000
Extracting function: _main from [16, 41]
allocated to 0x100154000
Relocation at '_main' + 16 from '_bar(Word1: 0x2d000000)
Resolving relocation at '_main' + 16 (0x100154010) from '_bar (0x100153000)(pcrel, type: 2, Size: 4).
loaded '_main' at: 0x100154000
65
$
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@129388 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-12 21:20:41 +00:00
|
|
|
// Flags == 0x8 marks a thumb function for ARM, which is fine as it
|
|
|
|
// doesn't require any special handling here.
|
|
|
|
if (STE->Flags != 0x0 && STE->Flags != 0x8)
|
2011-04-08 17:31:24 +00:00
|
|
|
return Error("unexpected symbol type!");
|
|
|
|
|
|
|
|
// Remember the symbol.
|
MCJIT lazy relocation resolution and symbol address re-assignment.
Add handling for tracking the relocations on symbols and resolving them.
Keep track of the relocations even after they are resolved so that if
the RuntimeDyld client moves the object, it can update the address and any
relocations to that object will be updated.
For our trival object file load/run test harness (llvm-rtdyld), this enables
relocations between functions located in the same object module. It should
be trivially extendable to load multiple objects with mutual references.
As a simple example, the following now works (running on x86_64 Darwin 10.6):
$ cat t.c
int bar() {
return 65;
}
int main() {
return bar();
}
$ clang t.c -fno-asynchronous-unwind-tables -o t.o -c
$ otool -vt t.o
t.o:
(__TEXT,__text) section
_bar:
0000000000000000 pushq %rbp
0000000000000001 movq %rsp,%rbp
0000000000000004 movl $0x00000041,%eax
0000000000000009 popq %rbp
000000000000000a ret
000000000000000b nopl 0x00(%rax,%rax)
_main:
0000000000000010 pushq %rbp
0000000000000011 movq %rsp,%rbp
0000000000000014 subq $0x10,%rsp
0000000000000018 movl $0x00000000,0xfc(%rbp)
000000000000001f callq 0x00000024
0000000000000024 addq $0x10,%rsp
0000000000000028 popq %rbp
0000000000000029 ret
$ llvm-rtdyld t.o -debug-only=dyld ; echo $?
Function sym: '_bar' @ 0
Function sym: '_main' @ 16
Extracting function: _bar from [0, 15]
allocated to 0x100153000
Extracting function: _main from [16, 41]
allocated to 0x100154000
Relocation at '_main' + 16 from '_bar(Word1: 0x2d000000)
Resolving relocation at '_main' + 16 (0x100154010) from '_bar (0x100153000)(pcrel, type: 2, Size: 4).
loaded '_main' at: 0x100154000
65
$
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@129388 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-12 21:20:41 +00:00
|
|
|
Symbols.push_back(SymbolEntry(STE->Value, Name));
|
2011-04-08 17:31:24 +00:00
|
|
|
|
MCJIT lazy relocation resolution and symbol address re-assignment.
Add handling for tracking the relocations on symbols and resolving them.
Keep track of the relocations even after they are resolved so that if
the RuntimeDyld client moves the object, it can update the address and any
relocations to that object will be updated.
For our trival object file load/run test harness (llvm-rtdyld), this enables
relocations between functions located in the same object module. It should
be trivially extendable to load multiple objects with mutual references.
As a simple example, the following now works (running on x86_64 Darwin 10.6):
$ cat t.c
int bar() {
return 65;
}
int main() {
return bar();
}
$ clang t.c -fno-asynchronous-unwind-tables -o t.o -c
$ otool -vt t.o
t.o:
(__TEXT,__text) section
_bar:
0000000000000000 pushq %rbp
0000000000000001 movq %rsp,%rbp
0000000000000004 movl $0x00000041,%eax
0000000000000009 popq %rbp
000000000000000a ret
000000000000000b nopl 0x00(%rax,%rax)
_main:
0000000000000010 pushq %rbp
0000000000000011 movq %rsp,%rbp
0000000000000014 subq $0x10,%rsp
0000000000000018 movl $0x00000000,0xfc(%rbp)
000000000000001f callq 0x00000024
0000000000000024 addq $0x10,%rsp
0000000000000028 popq %rbp
0000000000000029 ret
$ llvm-rtdyld t.o -debug-only=dyld ; echo $?
Function sym: '_bar' @ 0
Function sym: '_main' @ 16
Extracting function: _bar from [0, 15]
allocated to 0x100153000
Extracting function: _main from [16, 41]
allocated to 0x100154000
Relocation at '_main' + 16 from '_bar(Word1: 0x2d000000)
Resolving relocation at '_main' + 16 (0x100154010) from '_bar (0x100153000)(pcrel, type: 2, Size: 4).
loaded '_main' at: 0x100154000
65
$
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@129388 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-12 21:20:41 +00:00
|
|
|
DEBUG(dbgs() << "Function sym: '" << Name << "' @ " <<
|
|
|
|
(Sect->Address + STE->Value) << "\n");
|
2011-04-08 17:31:24 +00:00
|
|
|
}
|
MCJIT lazy relocation resolution and symbol address re-assignment.
Add handling for tracking the relocations on symbols and resolving them.
Keep track of the relocations even after they are resolved so that if
the RuntimeDyld client moves the object, it can update the address and any
relocations to that object will be updated.
For our trival object file load/run test harness (llvm-rtdyld), this enables
relocations between functions located in the same object module. It should
be trivially extendable to load multiple objects with mutual references.
As a simple example, the following now works (running on x86_64 Darwin 10.6):
$ cat t.c
int bar() {
return 65;
}
int main() {
return bar();
}
$ clang t.c -fno-asynchronous-unwind-tables -o t.o -c
$ otool -vt t.o
t.o:
(__TEXT,__text) section
_bar:
0000000000000000 pushq %rbp
0000000000000001 movq %rsp,%rbp
0000000000000004 movl $0x00000041,%eax
0000000000000009 popq %rbp
000000000000000a ret
000000000000000b nopl 0x00(%rax,%rax)
_main:
0000000000000010 pushq %rbp
0000000000000011 movq %rsp,%rbp
0000000000000014 subq $0x10,%rsp
0000000000000018 movl $0x00000000,0xfc(%rbp)
000000000000001f callq 0x00000024
0000000000000024 addq $0x10,%rsp
0000000000000028 popq %rbp
0000000000000029 ret
$ llvm-rtdyld t.o -debug-only=dyld ; echo $?
Function sym: '_bar' @ 0
Function sym: '_main' @ 16
Extracting function: _bar from [0, 15]
allocated to 0x100153000
Extracting function: _main from [16, 41]
allocated to 0x100154000
Relocation at '_main' + 16 from '_bar(Word1: 0x2d000000)
Resolving relocation at '_main' + 16 (0x100154010) from '_bar (0x100153000)(pcrel, type: 2, Size: 4).
loaded '_main' at: 0x100154000
65
$
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@129388 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-12 21:20:41 +00:00
|
|
|
// Sort the symbols by address, just in case they didn't come in that way.
|
2011-04-08 17:31:24 +00:00
|
|
|
array_pod_sort(Symbols.begin(), Symbols.end());
|
2011-03-23 23:35:17 +00:00
|
|
|
|
2011-04-08 17:31:24 +00:00
|
|
|
// Extract the function data.
|
MCJIT lazy relocation resolution and symbol address re-assignment.
Add handling for tracking the relocations on symbols and resolving them.
Keep track of the relocations even after they are resolved so that if
the RuntimeDyld client moves the object, it can update the address and any
relocations to that object will be updated.
For our trival object file load/run test harness (llvm-rtdyld), this enables
relocations between functions located in the same object module. It should
be trivially extendable to load multiple objects with mutual references.
As a simple example, the following now works (running on x86_64 Darwin 10.6):
$ cat t.c
int bar() {
return 65;
}
int main() {
return bar();
}
$ clang t.c -fno-asynchronous-unwind-tables -o t.o -c
$ otool -vt t.o
t.o:
(__TEXT,__text) section
_bar:
0000000000000000 pushq %rbp
0000000000000001 movq %rsp,%rbp
0000000000000004 movl $0x00000041,%eax
0000000000000009 popq %rbp
000000000000000a ret
000000000000000b nopl 0x00(%rax,%rax)
_main:
0000000000000010 pushq %rbp
0000000000000011 movq %rsp,%rbp
0000000000000014 subq $0x10,%rsp
0000000000000018 movl $0x00000000,0xfc(%rbp)
000000000000001f callq 0x00000024
0000000000000024 addq $0x10,%rsp
0000000000000028 popq %rbp
0000000000000029 ret
$ llvm-rtdyld t.o -debug-only=dyld ; echo $?
Function sym: '_bar' @ 0
Function sym: '_main' @ 16
Extracting function: _bar from [0, 15]
allocated to 0x100153000
Extracting function: _main from [16, 41]
allocated to 0x100154000
Relocation at '_main' + 16 from '_bar(Word1: 0x2d000000)
Resolving relocation at '_main' + 16 (0x100154010) from '_bar (0x100153000)(pcrel, type: 2, Size: 4).
loaded '_main' at: 0x100154000
65
$
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@129388 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-12 21:20:41 +00:00
|
|
|
uint8_t *Base = (uint8_t*)Obj->getData(SegmentLC->FileOffset,
|
|
|
|
SegmentLC->FileSize).data();
|
2011-04-08 17:31:24 +00:00
|
|
|
for (unsigned i = 0, e = Symbols.size() - 1; i != e; ++i) {
|
MCJIT lazy relocation resolution and symbol address re-assignment.
Add handling for tracking the relocations on symbols and resolving them.
Keep track of the relocations even after they are resolved so that if
the RuntimeDyld client moves the object, it can update the address and any
relocations to that object will be updated.
For our trival object file load/run test harness (llvm-rtdyld), this enables
relocations between functions located in the same object module. It should
be trivially extendable to load multiple objects with mutual references.
As a simple example, the following now works (running on x86_64 Darwin 10.6):
$ cat t.c
int bar() {
return 65;
}
int main() {
return bar();
}
$ clang t.c -fno-asynchronous-unwind-tables -o t.o -c
$ otool -vt t.o
t.o:
(__TEXT,__text) section
_bar:
0000000000000000 pushq %rbp
0000000000000001 movq %rsp,%rbp
0000000000000004 movl $0x00000041,%eax
0000000000000009 popq %rbp
000000000000000a ret
000000000000000b nopl 0x00(%rax,%rax)
_main:
0000000000000010 pushq %rbp
0000000000000011 movq %rsp,%rbp
0000000000000014 subq $0x10,%rsp
0000000000000018 movl $0x00000000,0xfc(%rbp)
000000000000001f callq 0x00000024
0000000000000024 addq $0x10,%rsp
0000000000000028 popq %rbp
0000000000000029 ret
$ llvm-rtdyld t.o -debug-only=dyld ; echo $?
Function sym: '_bar' @ 0
Function sym: '_main' @ 16
Extracting function: _bar from [0, 15]
allocated to 0x100153000
Extracting function: _main from [16, 41]
allocated to 0x100154000
Relocation at '_main' + 16 from '_bar(Word1: 0x2d000000)
Resolving relocation at '_main' + 16 (0x100154010) from '_bar (0x100153000)(pcrel, type: 2, Size: 4).
loaded '_main' at: 0x100154000
65
$
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@129388 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-12 21:20:41 +00:00
|
|
|
uint64_t StartOffset = Sect->Address + Symbols[i].first;
|
2011-04-08 17:31:24 +00:00
|
|
|
uint64_t EndOffset = Symbols[i + 1].first - 1;
|
|
|
|
DEBUG(dbgs() << "Extracting function: " << Symbols[i].second
|
|
|
|
<< " from [" << StartOffset << ", " << EndOffset << "]\n");
|
|
|
|
extractFunction(Symbols[i].second, Base + StartOffset, Base + EndOffset);
|
|
|
|
}
|
|
|
|
// The last symbol we do after since the end address is calculated
|
|
|
|
// differently because there is no next symbol to reference.
|
|
|
|
uint64_t StartOffset = Symbols[Symbols.size() - 1].first;
|
|
|
|
uint64_t EndOffset = Sect->Size - 1;
|
|
|
|
DEBUG(dbgs() << "Extracting function: " << Symbols[Symbols.size()-1].second
|
|
|
|
<< " from [" << StartOffset << ", " << EndOffset << "]\n");
|
|
|
|
extractFunction(Symbols[Symbols.size()-1].second,
|
|
|
|
Base + StartOffset, Base + EndOffset);
|
2011-03-22 01:06:42 +00:00
|
|
|
|
MCJIT lazy relocation resolution and symbol address re-assignment.
Add handling for tracking the relocations on symbols and resolving them.
Keep track of the relocations even after they are resolved so that if
the RuntimeDyld client moves the object, it can update the address and any
relocations to that object will be updated.
For our trival object file load/run test harness (llvm-rtdyld), this enables
relocations between functions located in the same object module. It should
be trivially extendable to load multiple objects with mutual references.
As a simple example, the following now works (running on x86_64 Darwin 10.6):
$ cat t.c
int bar() {
return 65;
}
int main() {
return bar();
}
$ clang t.c -fno-asynchronous-unwind-tables -o t.o -c
$ otool -vt t.o
t.o:
(__TEXT,__text) section
_bar:
0000000000000000 pushq %rbp
0000000000000001 movq %rsp,%rbp
0000000000000004 movl $0x00000041,%eax
0000000000000009 popq %rbp
000000000000000a ret
000000000000000b nopl 0x00(%rax,%rax)
_main:
0000000000000010 pushq %rbp
0000000000000011 movq %rsp,%rbp
0000000000000014 subq $0x10,%rsp
0000000000000018 movl $0x00000000,0xfc(%rbp)
000000000000001f callq 0x00000024
0000000000000024 addq $0x10,%rsp
0000000000000028 popq %rbp
0000000000000029 ret
$ llvm-rtdyld t.o -debug-only=dyld ; echo $?
Function sym: '_bar' @ 0
Function sym: '_main' @ 16
Extracting function: _bar from [0, 15]
allocated to 0x100153000
Extracting function: _main from [16, 41]
allocated to 0x100154000
Relocation at '_main' + 16 from '_bar(Word1: 0x2d000000)
Resolving relocation at '_main' + 16 (0x100154010) from '_bar (0x100153000)(pcrel, type: 2, Size: 4).
loaded '_main' at: 0x100154000
65
$
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@129388 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-12 21:20:41 +00:00
|
|
|
// Now extract the relocation information for each function and process it.
|
|
|
|
for (unsigned j = 0; j != Sect->NumRelocationTableEntries; ++j) {
|
|
|
|
InMemoryStruct<macho::RelocationEntry> RE;
|
|
|
|
Obj->ReadRelocationEntry(Sect->RelocationTableOffset, j, RE);
|
|
|
|
if (RE->Word0 & macho::RF_Scattered)
|
|
|
|
return Error("NOT YET IMPLEMENTED: scattered relocations.");
|
|
|
|
// Word0 of the relocation is the offset into the section where the
|
|
|
|
// relocation should be applied. We need to translate that into an
|
|
|
|
// offset into a function since that's our atom.
|
|
|
|
uint32_t Offset = RE->Word0;
|
|
|
|
// Look for the function containing the address. This is used for JIT
|
|
|
|
// code, so the number of functions in section is almost always going
|
|
|
|
// to be very small (usually just one), so until we have use cases
|
|
|
|
// where that's not true, just use a trivial linear search.
|
|
|
|
unsigned SymbolNum;
|
|
|
|
unsigned NumSymbols = Symbols.size();
|
|
|
|
assert(NumSymbols > 0 && Symbols[0].first <= Offset &&
|
|
|
|
"No symbol containing relocation!");
|
|
|
|
for (SymbolNum = 0; SymbolNum < NumSymbols - 1; ++SymbolNum)
|
|
|
|
if (Symbols[SymbolNum + 1].first > Offset)
|
|
|
|
break;
|
|
|
|
// Adjust the offset to be relative to the symbol.
|
|
|
|
Offset -= Symbols[SymbolNum].first;
|
|
|
|
// Get the name of the symbol containing the relocation.
|
|
|
|
StringRef TargetName = SymbolNames[SymbolNum];
|
|
|
|
|
|
|
|
bool isExtern = (RE->Word1 >> 27) & 1;
|
|
|
|
// Figure out the source symbol of the relocation. If isExtern is true,
|
|
|
|
// this relocation references the symbol table, otherwise it references
|
|
|
|
// a section in the same object, numbered from 1 through NumSections
|
|
|
|
// (SectionBases is [0, NumSections-1]).
|
|
|
|
// FIXME: Some targets (ARM) use internal relocations even for
|
|
|
|
// externally visible symbols, if the definition is in the same
|
|
|
|
// file as the reference. We need to convert those back to by-name
|
|
|
|
// references. We can resolve the address based on the section
|
|
|
|
// offset and see if we have a symbol at that address. If we do,
|
|
|
|
// use that; otherwise, puke.
|
|
|
|
if (!isExtern)
|
|
|
|
return Error("Internal relocations not supported.");
|
|
|
|
uint32_t SourceNum = RE->Word1 & 0xffffff; // 24-bit value
|
|
|
|
StringRef SourceName = SymbolNames[SourceNum];
|
|
|
|
|
|
|
|
// FIXME: Get the relocation addend from the target address.
|
|
|
|
|
|
|
|
// Now store the relocation information. Associate it with the source
|
|
|
|
// symbol.
|
|
|
|
Relocations[SourceName].push_back(RelocationEntry(TargetName,
|
|
|
|
Offset,
|
|
|
|
RE->Word1,
|
|
|
|
0 /*Addend*/));
|
|
|
|
DEBUG(dbgs() << "Relocation at '" << TargetName << "' + " << Offset
|
|
|
|
<< " from '" << SourceName << "(Word1: "
|
|
|
|
<< format("0x%x", RE->Word1) << ")\n");
|
|
|
|
}
|
|
|
|
}
|
2011-03-21 22:15:52 +00:00
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
bool RuntimeDyldImpl::
|
|
|
|
loadSegment64(const MachOObject *Obj,
|
|
|
|
const MachOObject::LoadCommandInfo *SegmentLCI,
|
|
|
|
const InMemoryStruct<macho::SymtabLoadCommand> &SymtabLC) {
|
|
|
|
InMemoryStruct<macho::Segment64LoadCommand> Segment64LC;
|
|
|
|
Obj->ReadSegment64LoadCommand(*SegmentLCI, Segment64LC);
|
|
|
|
if (!Segment64LC)
|
|
|
|
return Error("unable to load segment load command");
|
|
|
|
|
2011-04-06 01:11:05 +00:00
|
|
|
for (unsigned SectNum = 0; SectNum != Segment64LC->NumSections; ++SectNum) {
|
2011-03-21 22:15:52 +00:00
|
|
|
InMemoryStruct<macho::Section64> Sect;
|
2011-04-06 01:11:05 +00:00
|
|
|
Obj->ReadSection64(*SegmentLCI, SectNum, Sect);
|
2011-03-21 22:15:52 +00:00
|
|
|
if (!Sect)
|
2011-04-06 01:11:05 +00:00
|
|
|
return Error("unable to load section: '" + Twine(SectNum) + "'");
|
2011-03-21 22:15:52 +00:00
|
|
|
|
|
|
|
// FIXME: Improve check.
|
|
|
|
if (Sect->Flags != 0x80000400)
|
|
|
|
return Error("unsupported section type!");
|
|
|
|
|
2011-04-06 01:11:05 +00:00
|
|
|
// Address and names of symbols in the section.
|
|
|
|
typedef std::pair<uint64_t, StringRef> SymbolEntry;
|
|
|
|
SmallVector<SymbolEntry, 64> Symbols;
|
MCJIT lazy relocation resolution and symbol address re-assignment.
Add handling for tracking the relocations on symbols and resolving them.
Keep track of the relocations even after they are resolved so that if
the RuntimeDyld client moves the object, it can update the address and any
relocations to that object will be updated.
For our trival object file load/run test harness (llvm-rtdyld), this enables
relocations between functions located in the same object module. It should
be trivially extendable to load multiple objects with mutual references.
As a simple example, the following now works (running on x86_64 Darwin 10.6):
$ cat t.c
int bar() {
return 65;
}
int main() {
return bar();
}
$ clang t.c -fno-asynchronous-unwind-tables -o t.o -c
$ otool -vt t.o
t.o:
(__TEXT,__text) section
_bar:
0000000000000000 pushq %rbp
0000000000000001 movq %rsp,%rbp
0000000000000004 movl $0x00000041,%eax
0000000000000009 popq %rbp
000000000000000a ret
000000000000000b nopl 0x00(%rax,%rax)
_main:
0000000000000010 pushq %rbp
0000000000000011 movq %rsp,%rbp
0000000000000014 subq $0x10,%rsp
0000000000000018 movl $0x00000000,0xfc(%rbp)
000000000000001f callq 0x00000024
0000000000000024 addq $0x10,%rsp
0000000000000028 popq %rbp
0000000000000029 ret
$ llvm-rtdyld t.o -debug-only=dyld ; echo $?
Function sym: '_bar' @ 0
Function sym: '_main' @ 16
Extracting function: _bar from [0, 15]
allocated to 0x100153000
Extracting function: _main from [16, 41]
allocated to 0x100154000
Relocation at '_main' + 16 from '_bar(Word1: 0x2d000000)
Resolving relocation at '_main' + 16 (0x100154010) from '_bar (0x100153000)(pcrel, type: 2, Size: 4).
loaded '_main' at: 0x100154000
65
$
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@129388 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-12 21:20:41 +00:00
|
|
|
// Index of all the names, in this section or not. Used when we're
|
|
|
|
// dealing with relocation entries.
|
|
|
|
SmallVector<StringRef, 64> SymbolNames;
|
2011-04-06 01:11:05 +00:00
|
|
|
for (unsigned i = 0; i != SymtabLC->NumSymbolTableEntries; ++i) {
|
|
|
|
InMemoryStruct<macho::Symbol64TableEntry> STE;
|
|
|
|
Obj->ReadSymbol64TableEntry(SymtabLC->SymbolTableOffset, i, STE);
|
|
|
|
if (!STE)
|
|
|
|
return Error("unable to read symbol: '" + Twine(i) + "'");
|
|
|
|
if (STE->SectionIndex > Segment64LC->NumSections)
|
2011-04-09 10:10:35 +00:00
|
|
|
return Error("invalid section index for symbol: '" + Twine(i) + "'");
|
MCJIT lazy relocation resolution and symbol address re-assignment.
Add handling for tracking the relocations on symbols and resolving them.
Keep track of the relocations even after they are resolved so that if
the RuntimeDyld client moves the object, it can update the address and any
relocations to that object will be updated.
For our trival object file load/run test harness (llvm-rtdyld), this enables
relocations between functions located in the same object module. It should
be trivially extendable to load multiple objects with mutual references.
As a simple example, the following now works (running on x86_64 Darwin 10.6):
$ cat t.c
int bar() {
return 65;
}
int main() {
return bar();
}
$ clang t.c -fno-asynchronous-unwind-tables -o t.o -c
$ otool -vt t.o
t.o:
(__TEXT,__text) section
_bar:
0000000000000000 pushq %rbp
0000000000000001 movq %rsp,%rbp
0000000000000004 movl $0x00000041,%eax
0000000000000009 popq %rbp
000000000000000a ret
000000000000000b nopl 0x00(%rax,%rax)
_main:
0000000000000010 pushq %rbp
0000000000000011 movq %rsp,%rbp
0000000000000014 subq $0x10,%rsp
0000000000000018 movl $0x00000000,0xfc(%rbp)
000000000000001f callq 0x00000024
0000000000000024 addq $0x10,%rsp
0000000000000028 popq %rbp
0000000000000029 ret
$ llvm-rtdyld t.o -debug-only=dyld ; echo $?
Function sym: '_bar' @ 0
Function sym: '_main' @ 16
Extracting function: _bar from [0, 15]
allocated to 0x100153000
Extracting function: _main from [16, 41]
allocated to 0x100154000
Relocation at '_main' + 16 from '_bar(Word1: 0x2d000000)
Resolving relocation at '_main' + 16 (0x100154010) from '_bar (0x100153000)(pcrel, type: 2, Size: 4).
loaded '_main' at: 0x100154000
65
$
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@129388 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-12 21:20:41 +00:00
|
|
|
// Get the symbol name.
|
|
|
|
StringRef Name = Obj->getStringAtIndex(STE->StringIndex);
|
|
|
|
SymbolNames.push_back(Name);
|
2011-04-06 01:11:05 +00:00
|
|
|
|
|
|
|
// Just skip symbols not defined in this section.
|
2011-04-08 21:11:20 +00:00
|
|
|
if ((unsigned)STE->SectionIndex - 1 != SectNum)
|
2011-04-06 01:11:05 +00:00
|
|
|
continue;
|
|
|
|
|
|
|
|
// FIXME: Check the symbol type and flags.
|
|
|
|
if (STE->Type != 0xF) // external, defined in this section.
|
|
|
|
return Error("unexpected symbol type!");
|
|
|
|
if (STE->Flags != 0x0)
|
|
|
|
return Error("unexpected symbol type!");
|
|
|
|
|
|
|
|
// Remember the symbol.
|
MCJIT lazy relocation resolution and symbol address re-assignment.
Add handling for tracking the relocations on symbols and resolving them.
Keep track of the relocations even after they are resolved so that if
the RuntimeDyld client moves the object, it can update the address and any
relocations to that object will be updated.
For our trival object file load/run test harness (llvm-rtdyld), this enables
relocations between functions located in the same object module. It should
be trivially extendable to load multiple objects with mutual references.
As a simple example, the following now works (running on x86_64 Darwin 10.6):
$ cat t.c
int bar() {
return 65;
}
int main() {
return bar();
}
$ clang t.c -fno-asynchronous-unwind-tables -o t.o -c
$ otool -vt t.o
t.o:
(__TEXT,__text) section
_bar:
0000000000000000 pushq %rbp
0000000000000001 movq %rsp,%rbp
0000000000000004 movl $0x00000041,%eax
0000000000000009 popq %rbp
000000000000000a ret
000000000000000b nopl 0x00(%rax,%rax)
_main:
0000000000000010 pushq %rbp
0000000000000011 movq %rsp,%rbp
0000000000000014 subq $0x10,%rsp
0000000000000018 movl $0x00000000,0xfc(%rbp)
000000000000001f callq 0x00000024
0000000000000024 addq $0x10,%rsp
0000000000000028 popq %rbp
0000000000000029 ret
$ llvm-rtdyld t.o -debug-only=dyld ; echo $?
Function sym: '_bar' @ 0
Function sym: '_main' @ 16
Extracting function: _bar from [0, 15]
allocated to 0x100153000
Extracting function: _main from [16, 41]
allocated to 0x100154000
Relocation at '_main' + 16 from '_bar(Word1: 0x2d000000)
Resolving relocation at '_main' + 16 (0x100154010) from '_bar (0x100153000)(pcrel, type: 2, Size: 4).
loaded '_main' at: 0x100154000
65
$
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@129388 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-12 21:20:41 +00:00
|
|
|
Symbols.push_back(SymbolEntry(STE->Value, Name));
|
2011-04-06 01:11:05 +00:00
|
|
|
|
MCJIT lazy relocation resolution and symbol address re-assignment.
Add handling for tracking the relocations on symbols and resolving them.
Keep track of the relocations even after they are resolved so that if
the RuntimeDyld client moves the object, it can update the address and any
relocations to that object will be updated.
For our trival object file load/run test harness (llvm-rtdyld), this enables
relocations between functions located in the same object module. It should
be trivially extendable to load multiple objects with mutual references.
As a simple example, the following now works (running on x86_64 Darwin 10.6):
$ cat t.c
int bar() {
return 65;
}
int main() {
return bar();
}
$ clang t.c -fno-asynchronous-unwind-tables -o t.o -c
$ otool -vt t.o
t.o:
(__TEXT,__text) section
_bar:
0000000000000000 pushq %rbp
0000000000000001 movq %rsp,%rbp
0000000000000004 movl $0x00000041,%eax
0000000000000009 popq %rbp
000000000000000a ret
000000000000000b nopl 0x00(%rax,%rax)
_main:
0000000000000010 pushq %rbp
0000000000000011 movq %rsp,%rbp
0000000000000014 subq $0x10,%rsp
0000000000000018 movl $0x00000000,0xfc(%rbp)
000000000000001f callq 0x00000024
0000000000000024 addq $0x10,%rsp
0000000000000028 popq %rbp
0000000000000029 ret
$ llvm-rtdyld t.o -debug-only=dyld ; echo $?
Function sym: '_bar' @ 0
Function sym: '_main' @ 16
Extracting function: _bar from [0, 15]
allocated to 0x100153000
Extracting function: _main from [16, 41]
allocated to 0x100154000
Relocation at '_main' + 16 from '_bar(Word1: 0x2d000000)
Resolving relocation at '_main' + 16 (0x100154010) from '_bar (0x100153000)(pcrel, type: 2, Size: 4).
loaded '_main' at: 0x100154000
65
$
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@129388 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-12 21:20:41 +00:00
|
|
|
DEBUG(dbgs() << "Function sym: '" << Name << "' @ " <<
|
|
|
|
(Sect->Address + STE->Value) << "\n");
|
2011-04-06 01:11:05 +00:00
|
|
|
}
|
MCJIT lazy relocation resolution and symbol address re-assignment.
Add handling for tracking the relocations on symbols and resolving them.
Keep track of the relocations even after they are resolved so that if
the RuntimeDyld client moves the object, it can update the address and any
relocations to that object will be updated.
For our trival object file load/run test harness (llvm-rtdyld), this enables
relocations between functions located in the same object module. It should
be trivially extendable to load multiple objects with mutual references.
As a simple example, the following now works (running on x86_64 Darwin 10.6):
$ cat t.c
int bar() {
return 65;
}
int main() {
return bar();
}
$ clang t.c -fno-asynchronous-unwind-tables -o t.o -c
$ otool -vt t.o
t.o:
(__TEXT,__text) section
_bar:
0000000000000000 pushq %rbp
0000000000000001 movq %rsp,%rbp
0000000000000004 movl $0x00000041,%eax
0000000000000009 popq %rbp
000000000000000a ret
000000000000000b nopl 0x00(%rax,%rax)
_main:
0000000000000010 pushq %rbp
0000000000000011 movq %rsp,%rbp
0000000000000014 subq $0x10,%rsp
0000000000000018 movl $0x00000000,0xfc(%rbp)
000000000000001f callq 0x00000024
0000000000000024 addq $0x10,%rsp
0000000000000028 popq %rbp
0000000000000029 ret
$ llvm-rtdyld t.o -debug-only=dyld ; echo $?
Function sym: '_bar' @ 0
Function sym: '_main' @ 16
Extracting function: _bar from [0, 15]
allocated to 0x100153000
Extracting function: _main from [16, 41]
allocated to 0x100154000
Relocation at '_main' + 16 from '_bar(Word1: 0x2d000000)
Resolving relocation at '_main' + 16 (0x100154010) from '_bar (0x100153000)(pcrel, type: 2, Size: 4).
loaded '_main' at: 0x100154000
65
$
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@129388 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-12 21:20:41 +00:00
|
|
|
// Sort the symbols by address, just in case they didn't come in that way.
|
2011-04-06 01:11:05 +00:00
|
|
|
array_pod_sort(Symbols.begin(), Symbols.end());
|
|
|
|
|
|
|
|
// Extract the function data.
|
|
|
|
uint8_t *Base = (uint8_t*)Obj->getData(Segment64LC->FileOffset,
|
|
|
|
Segment64LC->FileSize).data();
|
|
|
|
for (unsigned i = 0, e = Symbols.size() - 1; i != e; ++i) {
|
MCJIT lazy relocation resolution and symbol address re-assignment.
Add handling for tracking the relocations on symbols and resolving them.
Keep track of the relocations even after they are resolved so that if
the RuntimeDyld client moves the object, it can update the address and any
relocations to that object will be updated.
For our trival object file load/run test harness (llvm-rtdyld), this enables
relocations between functions located in the same object module. It should
be trivially extendable to load multiple objects with mutual references.
As a simple example, the following now works (running on x86_64 Darwin 10.6):
$ cat t.c
int bar() {
return 65;
}
int main() {
return bar();
}
$ clang t.c -fno-asynchronous-unwind-tables -o t.o -c
$ otool -vt t.o
t.o:
(__TEXT,__text) section
_bar:
0000000000000000 pushq %rbp
0000000000000001 movq %rsp,%rbp
0000000000000004 movl $0x00000041,%eax
0000000000000009 popq %rbp
000000000000000a ret
000000000000000b nopl 0x00(%rax,%rax)
_main:
0000000000000010 pushq %rbp
0000000000000011 movq %rsp,%rbp
0000000000000014 subq $0x10,%rsp
0000000000000018 movl $0x00000000,0xfc(%rbp)
000000000000001f callq 0x00000024
0000000000000024 addq $0x10,%rsp
0000000000000028 popq %rbp
0000000000000029 ret
$ llvm-rtdyld t.o -debug-only=dyld ; echo $?
Function sym: '_bar' @ 0
Function sym: '_main' @ 16
Extracting function: _bar from [0, 15]
allocated to 0x100153000
Extracting function: _main from [16, 41]
allocated to 0x100154000
Relocation at '_main' + 16 from '_bar(Word1: 0x2d000000)
Resolving relocation at '_main' + 16 (0x100154010) from '_bar (0x100153000)(pcrel, type: 2, Size: 4).
loaded '_main' at: 0x100154000
65
$
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@129388 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-12 21:20:41 +00:00
|
|
|
uint64_t StartOffset = Sect->Address + Symbols[i].first;
|
2011-04-06 01:11:05 +00:00
|
|
|
uint64_t EndOffset = Symbols[i + 1].first - 1;
|
|
|
|
DEBUG(dbgs() << "Extracting function: " << Symbols[i].second
|
|
|
|
<< " from [" << StartOffset << ", " << EndOffset << "]\n");
|
|
|
|
extractFunction(Symbols[i].second, Base + StartOffset, Base + EndOffset);
|
|
|
|
}
|
|
|
|
// The last symbol we do after since the end address is calculated
|
|
|
|
// differently because there is no next symbol to reference.
|
|
|
|
uint64_t StartOffset = Symbols[Symbols.size() - 1].first;
|
|
|
|
uint64_t EndOffset = Sect->Size - 1;
|
|
|
|
DEBUG(dbgs() << "Extracting function: " << Symbols[Symbols.size()-1].second
|
|
|
|
<< " from [" << StartOffset << ", " << EndOffset << "]\n");
|
|
|
|
extractFunction(Symbols[Symbols.size()-1].second,
|
|
|
|
Base + StartOffset, Base + EndOffset);
|
2011-03-23 19:52:00 +00:00
|
|
|
|
MCJIT lazy relocation resolution and symbol address re-assignment.
Add handling for tracking the relocations on symbols and resolving them.
Keep track of the relocations even after they are resolved so that if
the RuntimeDyld client moves the object, it can update the address and any
relocations to that object will be updated.
For our trival object file load/run test harness (llvm-rtdyld), this enables
relocations between functions located in the same object module. It should
be trivially extendable to load multiple objects with mutual references.
As a simple example, the following now works (running on x86_64 Darwin 10.6):
$ cat t.c
int bar() {
return 65;
}
int main() {
return bar();
}
$ clang t.c -fno-asynchronous-unwind-tables -o t.o -c
$ otool -vt t.o
t.o:
(__TEXT,__text) section
_bar:
0000000000000000 pushq %rbp
0000000000000001 movq %rsp,%rbp
0000000000000004 movl $0x00000041,%eax
0000000000000009 popq %rbp
000000000000000a ret
000000000000000b nopl 0x00(%rax,%rax)
_main:
0000000000000010 pushq %rbp
0000000000000011 movq %rsp,%rbp
0000000000000014 subq $0x10,%rsp
0000000000000018 movl $0x00000000,0xfc(%rbp)
000000000000001f callq 0x00000024
0000000000000024 addq $0x10,%rsp
0000000000000028 popq %rbp
0000000000000029 ret
$ llvm-rtdyld t.o -debug-only=dyld ; echo $?
Function sym: '_bar' @ 0
Function sym: '_main' @ 16
Extracting function: _bar from [0, 15]
allocated to 0x100153000
Extracting function: _main from [16, 41]
allocated to 0x100154000
Relocation at '_main' + 16 from '_bar(Word1: 0x2d000000)
Resolving relocation at '_main' + 16 (0x100154010) from '_bar (0x100153000)(pcrel, type: 2, Size: 4).
loaded '_main' at: 0x100154000
65
$
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@129388 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-12 21:20:41 +00:00
|
|
|
// Now extract the relocation information for each function and process it.
|
|
|
|
for (unsigned j = 0; j != Sect->NumRelocationTableEntries; ++j) {
|
|
|
|
InMemoryStruct<macho::RelocationEntry> RE;
|
|
|
|
Obj->ReadRelocationEntry(Sect->RelocationTableOffset, j, RE);
|
|
|
|
if (RE->Word0 & macho::RF_Scattered)
|
|
|
|
return Error("NOT YET IMPLEMENTED: scattered relocations.");
|
|
|
|
// Word0 of the relocation is the offset into the section where the
|
|
|
|
// relocation should be applied. We need to translate that into an
|
|
|
|
// offset into a function since that's our atom.
|
|
|
|
uint32_t Offset = RE->Word0;
|
|
|
|
// Look for the function containing the address. This is used for JIT
|
|
|
|
// code, so the number of functions in section is almost always going
|
|
|
|
// to be very small (usually just one), so until we have use cases
|
|
|
|
// where that's not true, just use a trivial linear search.
|
|
|
|
unsigned SymbolNum;
|
|
|
|
unsigned NumSymbols = Symbols.size();
|
|
|
|
assert(NumSymbols > 0 && Symbols[0].first <= Offset &&
|
|
|
|
"No symbol containing relocation!");
|
|
|
|
for (SymbolNum = 0; SymbolNum < NumSymbols - 1; ++SymbolNum)
|
|
|
|
if (Symbols[SymbolNum + 1].first > Offset)
|
|
|
|
break;
|
|
|
|
// Adjust the offset to be relative to the symbol.
|
|
|
|
Offset -= Symbols[SymbolNum].first;
|
|
|
|
// Get the name of the symbol containing the relocation.
|
|
|
|
StringRef TargetName = SymbolNames[SymbolNum];
|
|
|
|
|
|
|
|
bool isExtern = (RE->Word1 >> 27) & 1;
|
|
|
|
// Figure out the source symbol of the relocation. If isExtern is true,
|
|
|
|
// this relocation references the symbol table, otherwise it references
|
|
|
|
// a section in the same object, numbered from 1 through NumSections
|
|
|
|
// (SectionBases is [0, NumSections-1]).
|
|
|
|
if (!isExtern)
|
|
|
|
return Error("Internal relocations not supported.");
|
|
|
|
uint32_t SourceNum = RE->Word1 & 0xffffff; // 24-bit value
|
|
|
|
StringRef SourceName = SymbolNames[SourceNum];
|
|
|
|
|
|
|
|
// FIXME: Get the relocation addend from the target address.
|
|
|
|
|
|
|
|
// Now store the relocation information. Associate it with the source
|
|
|
|
// symbol.
|
|
|
|
Relocations[SourceName].push_back(RelocationEntry(TargetName,
|
|
|
|
Offset,
|
|
|
|
RE->Word1,
|
|
|
|
0 /*Addend*/));
|
|
|
|
DEBUG(dbgs() << "Relocation at '" << TargetName << "' + " << Offset
|
|
|
|
<< " from '" << SourceName << "(Word1: "
|
|
|
|
<< format("0x%x", RE->Word1) << ")\n");
|
|
|
|
}
|
|
|
|
}
|
2011-03-21 22:15:52 +00:00
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
|
|
|
bool RuntimeDyldImpl::loadObject(MemoryBuffer *InputBuffer) {
|
|
|
|
// If the linker is in an error state, don't do anything.
|
|
|
|
if (hasError())
|
|
|
|
return true;
|
|
|
|
// Load the Mach-O wrapper object.
|
|
|
|
std::string ErrorStr;
|
|
|
|
OwningPtr<MachOObject> Obj(
|
|
|
|
MachOObject::LoadFromBuffer(InputBuffer, &ErrorStr));
|
|
|
|
if (!Obj)
|
|
|
|
return Error("unable to load object: '" + ErrorStr + "'");
|
|
|
|
|
2011-03-23 22:06:06 +00:00
|
|
|
// Get the CPU type information from the header.
|
|
|
|
const macho::Header &Header = Obj->getHeader();
|
|
|
|
|
|
|
|
// FIXME: Error checking that the loaded object is compatible with
|
|
|
|
// the system we're running on.
|
|
|
|
CPUType = Header.CPUType;
|
|
|
|
CPUSubtype = Header.CPUSubtype;
|
|
|
|
|
2011-03-21 22:15:52 +00:00
|
|
|
// Validate that the load commands match what we expect.
|
|
|
|
const MachOObject::LoadCommandInfo *SegmentLCI = 0, *SymtabLCI = 0,
|
|
|
|
*DysymtabLCI = 0;
|
2011-03-23 22:06:06 +00:00
|
|
|
for (unsigned i = 0; i != Header.NumLoadCommands; ++i) {
|
2011-03-21 22:15:52 +00:00
|
|
|
const MachOObject::LoadCommandInfo &LCI = Obj->getLoadCommandInfo(i);
|
|
|
|
switch (LCI.Command.Type) {
|
|
|
|
case macho::LCT_Segment:
|
|
|
|
case macho::LCT_Segment64:
|
|
|
|
if (SegmentLCI)
|
|
|
|
return Error("unexpected input object (multiple segments)");
|
|
|
|
SegmentLCI = &LCI;
|
|
|
|
break;
|
|
|
|
case macho::LCT_Symtab:
|
|
|
|
if (SymtabLCI)
|
|
|
|
return Error("unexpected input object (multiple symbol tables)");
|
|
|
|
SymtabLCI = &LCI;
|
|
|
|
break;
|
|
|
|
case macho::LCT_Dysymtab:
|
|
|
|
if (DysymtabLCI)
|
|
|
|
return Error("unexpected input object (multiple symbol tables)");
|
|
|
|
DysymtabLCI = &LCI;
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
return Error("unexpected input object (unexpected load command");
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!SymtabLCI)
|
|
|
|
return Error("no symbol table found in object");
|
|
|
|
if (!SegmentLCI)
|
|
|
|
return Error("no symbol table found in object");
|
|
|
|
|
|
|
|
// Read and register the symbol table data.
|
|
|
|
InMemoryStruct<macho::SymtabLoadCommand> SymtabLC;
|
|
|
|
Obj->ReadSymtabLoadCommand(*SymtabLCI, SymtabLC);
|
|
|
|
if (!SymtabLC)
|
|
|
|
return Error("unable to load symbol table load command");
|
|
|
|
Obj->RegisterStringTable(*SymtabLC);
|
|
|
|
|
|
|
|
// Read the dynamic link-edit information, if present (not present in static
|
|
|
|
// objects).
|
|
|
|
if (DysymtabLCI) {
|
|
|
|
InMemoryStruct<macho::DysymtabLoadCommand> DysymtabLC;
|
|
|
|
Obj->ReadDysymtabLoadCommand(*DysymtabLCI, DysymtabLC);
|
|
|
|
if (!DysymtabLC)
|
|
|
|
return Error("unable to load dynamic link-exit load command");
|
|
|
|
|
|
|
|
// FIXME: We don't support anything interesting yet.
|
2011-03-23 19:52:00 +00:00
|
|
|
// if (DysymtabLC->LocalSymbolsIndex != 0)
|
|
|
|
// return Error("NOT YET IMPLEMENTED: local symbol entries");
|
|
|
|
// if (DysymtabLC->ExternalSymbolsIndex != 0)
|
|
|
|
// return Error("NOT YET IMPLEMENTED: non-external symbol entries");
|
|
|
|
// if (DysymtabLC->UndefinedSymbolsIndex != SymtabLC->NumSymbolTableEntries)
|
|
|
|
// return Error("NOT YET IMPLEMENTED: undefined symbol entries");
|
2011-03-21 22:15:52 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
// Load the segment load command.
|
|
|
|
if (SegmentLCI->Command.Type == macho::LCT_Segment) {
|
|
|
|
if (loadSegment32(Obj.get(), SegmentLCI, SymtabLC))
|
|
|
|
return true;
|
|
|
|
} else {
|
|
|
|
if (loadSegment64(Obj.get(), SegmentLCI, SymtabLC))
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
|
MCJIT lazy relocation resolution and symbol address re-assignment.
Add handling for tracking the relocations on symbols and resolving them.
Keep track of the relocations even after they are resolved so that if
the RuntimeDyld client moves the object, it can update the address and any
relocations to that object will be updated.
For our trival object file load/run test harness (llvm-rtdyld), this enables
relocations between functions located in the same object module. It should
be trivially extendable to load multiple objects with mutual references.
As a simple example, the following now works (running on x86_64 Darwin 10.6):
$ cat t.c
int bar() {
return 65;
}
int main() {
return bar();
}
$ clang t.c -fno-asynchronous-unwind-tables -o t.o -c
$ otool -vt t.o
t.o:
(__TEXT,__text) section
_bar:
0000000000000000 pushq %rbp
0000000000000001 movq %rsp,%rbp
0000000000000004 movl $0x00000041,%eax
0000000000000009 popq %rbp
000000000000000a ret
000000000000000b nopl 0x00(%rax,%rax)
_main:
0000000000000010 pushq %rbp
0000000000000011 movq %rsp,%rbp
0000000000000014 subq $0x10,%rsp
0000000000000018 movl $0x00000000,0xfc(%rbp)
000000000000001f callq 0x00000024
0000000000000024 addq $0x10,%rsp
0000000000000028 popq %rbp
0000000000000029 ret
$ llvm-rtdyld t.o -debug-only=dyld ; echo $?
Function sym: '_bar' @ 0
Function sym: '_main' @ 16
Extracting function: _bar from [0, 15]
allocated to 0x100153000
Extracting function: _main from [16, 41]
allocated to 0x100154000
Relocation at '_main' + 16 from '_bar(Word1: 0x2d000000)
Resolving relocation at '_main' + 16 (0x100154010) from '_bar (0x100153000)(pcrel, type: 2, Size: 4).
loaded '_main' at: 0x100154000
65
$
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@129388 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-12 21:20:41 +00:00
|
|
|
// Resolve the relocations for all symbols we currently know about.
|
|
|
|
void RuntimeDyldImpl::resolveRelocations() {
|
|
|
|
// Just iterate over the symbols in our symbol table and assign their
|
|
|
|
// addresses.
|
|
|
|
StringMap<uint8_t*>::iterator i = SymbolTable.begin();
|
|
|
|
StringMap<uint8_t*>::iterator e = SymbolTable.end();
|
|
|
|
for (;i != e; ++i)
|
|
|
|
reassignSymbolAddress(i->getKey(), i->getValue());
|
|
|
|
}
|
|
|
|
|
|
|
|
// Assign an address to a symbol name and resolve all the relocations
|
|
|
|
// associated with it.
|
|
|
|
void RuntimeDyldImpl::reassignSymbolAddress(StringRef Name, uint8_t *Addr) {
|
|
|
|
// Assign the address in our symbol table.
|
|
|
|
SymbolTable[Name] = Addr;
|
|
|
|
|
|
|
|
RelocationList &Relocs = Relocations[Name];
|
|
|
|
for (unsigned i = 0, e = Relocs.size(); i != e; ++i) {
|
|
|
|
RelocationEntry &RE = Relocs[i];
|
|
|
|
uint8_t *Target = SymbolTable[RE.Target] + RE.Offset;
|
|
|
|
bool isPCRel = (RE.Data >> 24) & 1;
|
|
|
|
unsigned Type = (RE.Data >> 28) & 0xf;
|
|
|
|
unsigned Size = 1 << ((RE.Data >> 25) & 3);
|
|
|
|
|
|
|
|
DEBUG(dbgs() << "Resolving relocation at '" << RE.Target
|
|
|
|
<< "' + " << RE.Offset << " (" << format("%p", Target) << ")"
|
|
|
|
<< " from '" << Name << " (" << format("%p", Addr) << ")"
|
|
|
|
<< "(" << (isPCRel ? "pcrel" : "absolute")
|
|
|
|
<< ", type: " << Type << ", Size: " << Size << ").\n");
|
|
|
|
|
|
|
|
resolveRelocation(Target, Addr, isPCRel, Type, Size);
|
|
|
|
RE.isResolved = true;
|
|
|
|
}
|
|
|
|
}
|
2011-03-21 22:15:52 +00:00
|
|
|
|
|
|
|
//===----------------------------------------------------------------------===//
|
|
|
|
// RuntimeDyld class implementation
|
2011-04-04 23:04:39 +00:00
|
|
|
RuntimeDyld::RuntimeDyld(RTDyldMemoryManager *MM) {
|
|
|
|
Dyld = new RuntimeDyldImpl(MM);
|
2011-03-21 22:15:52 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
RuntimeDyld::~RuntimeDyld() {
|
|
|
|
delete Dyld;
|
|
|
|
}
|
|
|
|
|
|
|
|
bool RuntimeDyld::loadObject(MemoryBuffer *InputBuffer) {
|
|
|
|
return Dyld->loadObject(InputBuffer);
|
|
|
|
}
|
|
|
|
|
2011-04-08 17:31:24 +00:00
|
|
|
void *RuntimeDyld::getSymbolAddress(StringRef Name) {
|
2011-03-21 22:15:52 +00:00
|
|
|
return Dyld->getSymbolAddress(Name);
|
|
|
|
}
|
|
|
|
|
MCJIT lazy relocation resolution and symbol address re-assignment.
Add handling for tracking the relocations on symbols and resolving them.
Keep track of the relocations even after they are resolved so that if
the RuntimeDyld client moves the object, it can update the address and any
relocations to that object will be updated.
For our trival object file load/run test harness (llvm-rtdyld), this enables
relocations between functions located in the same object module. It should
be trivially extendable to load multiple objects with mutual references.
As a simple example, the following now works (running on x86_64 Darwin 10.6):
$ cat t.c
int bar() {
return 65;
}
int main() {
return bar();
}
$ clang t.c -fno-asynchronous-unwind-tables -o t.o -c
$ otool -vt t.o
t.o:
(__TEXT,__text) section
_bar:
0000000000000000 pushq %rbp
0000000000000001 movq %rsp,%rbp
0000000000000004 movl $0x00000041,%eax
0000000000000009 popq %rbp
000000000000000a ret
000000000000000b nopl 0x00(%rax,%rax)
_main:
0000000000000010 pushq %rbp
0000000000000011 movq %rsp,%rbp
0000000000000014 subq $0x10,%rsp
0000000000000018 movl $0x00000000,0xfc(%rbp)
000000000000001f callq 0x00000024
0000000000000024 addq $0x10,%rsp
0000000000000028 popq %rbp
0000000000000029 ret
$ llvm-rtdyld t.o -debug-only=dyld ; echo $?
Function sym: '_bar' @ 0
Function sym: '_main' @ 16
Extracting function: _bar from [0, 15]
allocated to 0x100153000
Extracting function: _main from [16, 41]
allocated to 0x100154000
Relocation at '_main' + 16 from '_bar(Word1: 0x2d000000)
Resolving relocation at '_main' + 16 (0x100154010) from '_bar (0x100153000)(pcrel, type: 2, Size: 4).
loaded '_main' at: 0x100154000
65
$
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@129388 91177308-0d34-0410-b5e6-96231b3b80d8
2011-04-12 21:20:41 +00:00
|
|
|
void RuntimeDyld::resolveRelocations() {
|
|
|
|
Dyld->resolveRelocations();
|
|
|
|
}
|
|
|
|
|
|
|
|
void RuntimeDyld::reassignSymbolAddress(StringRef Name, uint8_t *Addr) {
|
|
|
|
Dyld->reassignSymbolAddress(Name, Addr);
|
|
|
|
}
|
|
|
|
|
2011-03-22 18:22:27 +00:00
|
|
|
StringRef RuntimeDyld::getErrorString() {
|
2011-03-22 18:19:42 +00:00
|
|
|
return Dyld->getErrorString();
|
|
|
|
}
|
|
|
|
|
2011-03-21 22:15:52 +00:00
|
|
|
} // end namespace llvm
|