Mark a CIE with a S in its augmentation string as signal frame.
This allows the code to properly handle signal frame data if none
of the signal frame pattern matchers work.
For a signal frame, DwarfSectionImpl<AddressType>::Eval needs to
continue the unwinding even if PC is zero. A zero PC means that the
program has crashed, and we should try to recover the real PC using the
return address on the stack or LR. This behavior is tested by
UnwindOffline.signal_{x86,x86_64}, which modify the libc.so files
so that the signal frame pattern matcher fails and the CIE/FDE
data is used instead.
Test: libunwindstack_test
Change-Id: I4655b070028fd984345311a5e743796f8c30ed36
830 lines
26 KiB
C++
830 lines
26 KiB
C++
/*
|
|
* Copyright (C) 2017 The Android Open Source Project
|
|
*
|
|
* Licensed under the Apache License, Version 2.0 (the "License");
|
|
* you may not use this file except in compliance with the License.
|
|
* You may obtain a copy of the License at
|
|
*
|
|
* http://www.apache.org/licenses/LICENSE-2.0
|
|
*
|
|
* Unless required by applicable law or agreed to in writing, software
|
|
* distributed under the License is distributed on an "AS IS" BASIS,
|
|
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
* See the License for the specific language governing permissions and
|
|
* limitations under the License.
|
|
*/
|
|
|
|
#include <stdint.h>
|
|
|
|
#include <unwindstack/DwarfError.h>
|
|
#include <unwindstack/DwarfLocation.h>
|
|
#include <unwindstack/DwarfMemory.h>
|
|
#include <unwindstack/DwarfSection.h>
|
|
#include <unwindstack/DwarfStructs.h>
|
|
#include <unwindstack/Elf.h>
|
|
#include <unwindstack/Log.h>
|
|
#include <unwindstack/Memory.h>
|
|
#include <unwindstack/Regs.h>
|
|
|
|
#include "DwarfCfa.h"
|
|
#include "DwarfDebugFrame.h"
|
|
#include "DwarfEhFrame.h"
|
|
#include "DwarfEncoding.h"
|
|
#include "DwarfOp.h"
|
|
#include "RegsInfo.h"
|
|
|
|
namespace unwindstack {
|
|
|
|
DwarfSection::DwarfSection(Memory* memory) : memory_(memory) {}
|
|
|
|
bool DwarfSection::Step(uint64_t pc, Regs* regs, Memory* process_memory, bool* finished,
|
|
bool* is_signal_frame) {
|
|
// Lookup the pc in the cache.
|
|
auto it = loc_regs_.upper_bound(pc);
|
|
if (it == loc_regs_.end() || pc < it->second.pc_start) {
|
|
last_error_.code = DWARF_ERROR_NONE;
|
|
const DwarfFde* fde = GetFdeFromPc(pc);
|
|
if (fde == nullptr || fde->cie == nullptr) {
|
|
last_error_.code = DWARF_ERROR_ILLEGAL_STATE;
|
|
return false;
|
|
}
|
|
|
|
// Now get the location information for this pc.
|
|
dwarf_loc_regs_t loc_regs;
|
|
if (!GetCfaLocationInfo(pc, fde, &loc_regs, regs->Arch())) {
|
|
return false;
|
|
}
|
|
loc_regs.cie = fde->cie;
|
|
|
|
// Store it in the cache.
|
|
it = loc_regs_.emplace(loc_regs.pc_end, std::move(loc_regs)).first;
|
|
}
|
|
|
|
*is_signal_frame = it->second.cie->is_signal_frame;
|
|
|
|
// Now eval the actual registers.
|
|
return Eval(it->second.cie, process_memory, it->second, regs, finished);
|
|
}
|
|
|
|
template <typename AddressType>
|
|
const DwarfCie* DwarfSectionImpl<AddressType>::GetCieFromOffset(uint64_t offset) {
|
|
auto cie_entry = cie_entries_.find(offset);
|
|
if (cie_entry != cie_entries_.end()) {
|
|
return &cie_entry->second;
|
|
}
|
|
DwarfCie* cie = &cie_entries_[offset];
|
|
memory_.set_data_offset(entries_offset_);
|
|
memory_.set_cur_offset(offset);
|
|
if (!FillInCieHeader(cie) || !FillInCie(cie)) {
|
|
// Erase the cached entry.
|
|
cie_entries_.erase(offset);
|
|
return nullptr;
|
|
}
|
|
return cie;
|
|
}
|
|
|
|
template <typename AddressType>
|
|
bool DwarfSectionImpl<AddressType>::FillInCieHeader(DwarfCie* cie) {
|
|
cie->lsda_encoding = DW_EH_PE_omit;
|
|
uint32_t length32;
|
|
if (!memory_.ReadBytes(&length32, sizeof(length32))) {
|
|
last_error_.code = DWARF_ERROR_MEMORY_INVALID;
|
|
last_error_.address = memory_.cur_offset();
|
|
return false;
|
|
}
|
|
if (length32 == static_cast<uint32_t>(-1)) {
|
|
// 64 bit Cie
|
|
uint64_t length64;
|
|
if (!memory_.ReadBytes(&length64, sizeof(length64))) {
|
|
last_error_.code = DWARF_ERROR_MEMORY_INVALID;
|
|
last_error_.address = memory_.cur_offset();
|
|
return false;
|
|
}
|
|
|
|
cie->cfa_instructions_end = memory_.cur_offset() + length64;
|
|
cie->fde_address_encoding = DW_EH_PE_sdata8;
|
|
|
|
uint64_t cie_id;
|
|
if (!memory_.ReadBytes(&cie_id, sizeof(cie_id))) {
|
|
last_error_.code = DWARF_ERROR_MEMORY_INVALID;
|
|
last_error_.address = memory_.cur_offset();
|
|
return false;
|
|
}
|
|
if (cie_id != cie64_value_) {
|
|
// This is not a Cie, something has gone horribly wrong.
|
|
last_error_.code = DWARF_ERROR_ILLEGAL_VALUE;
|
|
return false;
|
|
}
|
|
} else {
|
|
// 32 bit Cie
|
|
cie->cfa_instructions_end = memory_.cur_offset() + length32;
|
|
cie->fde_address_encoding = DW_EH_PE_sdata4;
|
|
|
|
uint32_t cie_id;
|
|
if (!memory_.ReadBytes(&cie_id, sizeof(cie_id))) {
|
|
last_error_.code = DWARF_ERROR_MEMORY_INVALID;
|
|
last_error_.address = memory_.cur_offset();
|
|
return false;
|
|
}
|
|
if (cie_id != cie32_value_) {
|
|
// This is not a Cie, something has gone horribly wrong.
|
|
last_error_.code = DWARF_ERROR_ILLEGAL_VALUE;
|
|
return false;
|
|
}
|
|
}
|
|
return true;
|
|
}
|
|
|
|
template <typename AddressType>
|
|
bool DwarfSectionImpl<AddressType>::FillInCie(DwarfCie* cie) {
|
|
if (!memory_.ReadBytes(&cie->version, sizeof(cie->version))) {
|
|
last_error_.code = DWARF_ERROR_MEMORY_INVALID;
|
|
last_error_.address = memory_.cur_offset();
|
|
return false;
|
|
}
|
|
|
|
if (cie->version != 1 && cie->version != 3 && cie->version != 4 && cie->version != 5) {
|
|
// Unrecognized version.
|
|
last_error_.code = DWARF_ERROR_UNSUPPORTED_VERSION;
|
|
return false;
|
|
}
|
|
|
|
// Read the augmentation string.
|
|
char aug_value;
|
|
do {
|
|
if (!memory_.ReadBytes(&aug_value, 1)) {
|
|
last_error_.code = DWARF_ERROR_MEMORY_INVALID;
|
|
last_error_.address = memory_.cur_offset();
|
|
return false;
|
|
}
|
|
cie->augmentation_string.push_back(aug_value);
|
|
} while (aug_value != '\0');
|
|
|
|
if (cie->version == 4 || cie->version == 5) {
|
|
// Skip the Address Size field since we only use it for validation.
|
|
memory_.set_cur_offset(memory_.cur_offset() + 1);
|
|
|
|
// Segment Size
|
|
if (!memory_.ReadBytes(&cie->segment_size, 1)) {
|
|
last_error_.code = DWARF_ERROR_MEMORY_INVALID;
|
|
last_error_.address = memory_.cur_offset();
|
|
return false;
|
|
}
|
|
}
|
|
|
|
// Code Alignment Factor
|
|
if (!memory_.ReadULEB128(&cie->code_alignment_factor)) {
|
|
last_error_.code = DWARF_ERROR_MEMORY_INVALID;
|
|
last_error_.address = memory_.cur_offset();
|
|
return false;
|
|
}
|
|
|
|
// Data Alignment Factor
|
|
if (!memory_.ReadSLEB128(&cie->data_alignment_factor)) {
|
|
last_error_.code = DWARF_ERROR_MEMORY_INVALID;
|
|
last_error_.address = memory_.cur_offset();
|
|
return false;
|
|
}
|
|
|
|
if (cie->version == 1) {
|
|
// Return Address is a single byte.
|
|
uint8_t return_address_register;
|
|
if (!memory_.ReadBytes(&return_address_register, 1)) {
|
|
last_error_.code = DWARF_ERROR_MEMORY_INVALID;
|
|
last_error_.address = memory_.cur_offset();
|
|
return false;
|
|
}
|
|
cie->return_address_register = return_address_register;
|
|
} else if (!memory_.ReadULEB128(&cie->return_address_register)) {
|
|
last_error_.code = DWARF_ERROR_MEMORY_INVALID;
|
|
last_error_.address = memory_.cur_offset();
|
|
return false;
|
|
}
|
|
|
|
if (cie->augmentation_string[0] != 'z') {
|
|
cie->cfa_instructions_offset = memory_.cur_offset();
|
|
return true;
|
|
}
|
|
|
|
uint64_t aug_length;
|
|
if (!memory_.ReadULEB128(&aug_length)) {
|
|
last_error_.code = DWARF_ERROR_MEMORY_INVALID;
|
|
last_error_.address = memory_.cur_offset();
|
|
return false;
|
|
}
|
|
cie->cfa_instructions_offset = memory_.cur_offset() + aug_length;
|
|
|
|
for (size_t i = 1; i < cie->augmentation_string.size(); i++) {
|
|
switch (cie->augmentation_string[i]) {
|
|
case 'L':
|
|
if (!memory_.ReadBytes(&cie->lsda_encoding, 1)) {
|
|
last_error_.code = DWARF_ERROR_MEMORY_INVALID;
|
|
last_error_.address = memory_.cur_offset();
|
|
return false;
|
|
}
|
|
break;
|
|
case 'P': {
|
|
uint8_t encoding;
|
|
if (!memory_.ReadBytes(&encoding, 1)) {
|
|
last_error_.code = DWARF_ERROR_MEMORY_INVALID;
|
|
last_error_.address = memory_.cur_offset();
|
|
return false;
|
|
}
|
|
memory_.set_pc_offset(pc_offset_);
|
|
if (!memory_.ReadEncodedValue<AddressType>(encoding, &cie->personality_handler)) {
|
|
last_error_.code = DWARF_ERROR_MEMORY_INVALID;
|
|
last_error_.address = memory_.cur_offset();
|
|
return false;
|
|
}
|
|
} break;
|
|
case 'R':
|
|
if (!memory_.ReadBytes(&cie->fde_address_encoding, 1)) {
|
|
last_error_.code = DWARF_ERROR_MEMORY_INVALID;
|
|
last_error_.address = memory_.cur_offset();
|
|
return false;
|
|
}
|
|
break;
|
|
case 'S':
|
|
cie->is_signal_frame = true;
|
|
break;
|
|
}
|
|
}
|
|
return true;
|
|
}
|
|
|
|
template <typename AddressType>
|
|
const DwarfFde* DwarfSectionImpl<AddressType>::GetFdeFromOffset(uint64_t offset) {
|
|
auto fde_entry = fde_entries_.find(offset);
|
|
if (fde_entry != fde_entries_.end()) {
|
|
return &fde_entry->second;
|
|
}
|
|
DwarfFde* fde = &fde_entries_[offset];
|
|
memory_.set_data_offset(entries_offset_);
|
|
memory_.set_cur_offset(offset);
|
|
if (!FillInFdeHeader(fde) || !FillInFde(fde)) {
|
|
fde_entries_.erase(offset);
|
|
return nullptr;
|
|
}
|
|
return fde;
|
|
}
|
|
|
|
template <typename AddressType>
|
|
bool DwarfSectionImpl<AddressType>::FillInFdeHeader(DwarfFde* fde) {
|
|
uint32_t length32;
|
|
if (!memory_.ReadBytes(&length32, sizeof(length32))) {
|
|
last_error_.code = DWARF_ERROR_MEMORY_INVALID;
|
|
last_error_.address = memory_.cur_offset();
|
|
return false;
|
|
}
|
|
|
|
if (length32 == static_cast<uint32_t>(-1)) {
|
|
// 64 bit Fde.
|
|
uint64_t length64;
|
|
if (!memory_.ReadBytes(&length64, sizeof(length64))) {
|
|
last_error_.code = DWARF_ERROR_MEMORY_INVALID;
|
|
last_error_.address = memory_.cur_offset();
|
|
return false;
|
|
}
|
|
fde->cfa_instructions_end = memory_.cur_offset() + length64;
|
|
|
|
uint64_t value64;
|
|
if (!memory_.ReadBytes(&value64, sizeof(value64))) {
|
|
last_error_.code = DWARF_ERROR_MEMORY_INVALID;
|
|
last_error_.address = memory_.cur_offset();
|
|
return false;
|
|
}
|
|
if (value64 == cie64_value_) {
|
|
// This is a Cie, this means something has gone wrong.
|
|
last_error_.code = DWARF_ERROR_ILLEGAL_VALUE;
|
|
return false;
|
|
}
|
|
|
|
// Get the Cie pointer, which is necessary to properly read the rest of
|
|
// of the Fde information.
|
|
fde->cie_offset = GetCieOffsetFromFde64(value64);
|
|
} else {
|
|
// 32 bit Fde.
|
|
fde->cfa_instructions_end = memory_.cur_offset() + length32;
|
|
|
|
uint32_t value32;
|
|
if (!memory_.ReadBytes(&value32, sizeof(value32))) {
|
|
last_error_.code = DWARF_ERROR_MEMORY_INVALID;
|
|
last_error_.address = memory_.cur_offset();
|
|
return false;
|
|
}
|
|
if (value32 == cie32_value_) {
|
|
// This is a Cie, this means something has gone wrong.
|
|
last_error_.code = DWARF_ERROR_ILLEGAL_VALUE;
|
|
return false;
|
|
}
|
|
|
|
// Get the Cie pointer, which is necessary to properly read the rest of
|
|
// of the Fde information.
|
|
fde->cie_offset = GetCieOffsetFromFde32(value32);
|
|
}
|
|
return true;
|
|
}
|
|
|
|
template <typename AddressType>
|
|
bool DwarfSectionImpl<AddressType>::FillInFde(DwarfFde* fde) {
|
|
uint64_t cur_offset = memory_.cur_offset();
|
|
|
|
const DwarfCie* cie = GetCieFromOffset(fde->cie_offset);
|
|
if (cie == nullptr) {
|
|
return false;
|
|
}
|
|
fde->cie = cie;
|
|
|
|
if (cie->segment_size != 0) {
|
|
// Skip over the segment selector for now.
|
|
cur_offset += cie->segment_size;
|
|
}
|
|
memory_.set_cur_offset(cur_offset);
|
|
|
|
// The load bias only applies to the start.
|
|
memory_.set_pc_offset(section_bias_);
|
|
bool valid = memory_.ReadEncodedValue<AddressType>(cie->fde_address_encoding, &fde->pc_start);
|
|
fde->pc_start = AdjustPcFromFde(fde->pc_start);
|
|
|
|
memory_.set_pc_offset(0);
|
|
if (!valid || !memory_.ReadEncodedValue<AddressType>(cie->fde_address_encoding, &fde->pc_end)) {
|
|
last_error_.code = DWARF_ERROR_MEMORY_INVALID;
|
|
last_error_.address = memory_.cur_offset();
|
|
return false;
|
|
}
|
|
fde->pc_end += fde->pc_start;
|
|
|
|
if (cie->augmentation_string.size() > 0 && cie->augmentation_string[0] == 'z') {
|
|
// Augmentation Size
|
|
uint64_t aug_length;
|
|
if (!memory_.ReadULEB128(&aug_length)) {
|
|
last_error_.code = DWARF_ERROR_MEMORY_INVALID;
|
|
last_error_.address = memory_.cur_offset();
|
|
return false;
|
|
}
|
|
uint64_t cur_offset = memory_.cur_offset();
|
|
|
|
memory_.set_pc_offset(pc_offset_);
|
|
if (!memory_.ReadEncodedValue<AddressType>(cie->lsda_encoding, &fde->lsda_address)) {
|
|
last_error_.code = DWARF_ERROR_MEMORY_INVALID;
|
|
last_error_.address = memory_.cur_offset();
|
|
return false;
|
|
}
|
|
|
|
// Set our position to after all of the augmentation data.
|
|
memory_.set_cur_offset(cur_offset + aug_length);
|
|
}
|
|
fde->cfa_instructions_offset = memory_.cur_offset();
|
|
|
|
return true;
|
|
}
|
|
|
|
template <typename AddressType>
|
|
bool DwarfSectionImpl<AddressType>::EvalExpression(const DwarfLocation& loc, Memory* regular_memory,
|
|
AddressType* value,
|
|
RegsInfo<AddressType>* regs_info,
|
|
bool* is_dex_pc) {
|
|
DwarfOp<AddressType> op(&memory_, regular_memory);
|
|
op.set_regs_info(regs_info);
|
|
|
|
// Need to evaluate the op data.
|
|
uint64_t end = loc.values[1];
|
|
uint64_t start = end - loc.values[0];
|
|
if (!op.Eval(start, end)) {
|
|
last_error_ = op.last_error();
|
|
return false;
|
|
}
|
|
if (op.StackSize() == 0) {
|
|
last_error_.code = DWARF_ERROR_ILLEGAL_STATE;
|
|
return false;
|
|
}
|
|
// We don't support an expression that evaluates to a register number.
|
|
if (op.is_register()) {
|
|
last_error_.code = DWARF_ERROR_NOT_IMPLEMENTED;
|
|
return false;
|
|
}
|
|
*value = op.StackAt(0);
|
|
if (is_dex_pc != nullptr && op.dex_pc_set()) {
|
|
*is_dex_pc = true;
|
|
}
|
|
return true;
|
|
}
|
|
|
|
template <typename AddressType>
|
|
struct EvalInfo {
|
|
const dwarf_loc_regs_t* loc_regs;
|
|
const DwarfCie* cie;
|
|
Memory* regular_memory;
|
|
AddressType cfa;
|
|
bool return_address_undefined = false;
|
|
RegsInfo<AddressType> regs_info;
|
|
};
|
|
|
|
template <typename AddressType>
|
|
bool DwarfSectionImpl<AddressType>::EvalRegister(const DwarfLocation* loc, uint32_t reg,
|
|
AddressType* reg_ptr, void* info) {
|
|
EvalInfo<AddressType>* eval_info = reinterpret_cast<EvalInfo<AddressType>*>(info);
|
|
Memory* regular_memory = eval_info->regular_memory;
|
|
switch (loc->type) {
|
|
case DWARF_LOCATION_OFFSET:
|
|
if (!regular_memory->ReadFully(eval_info->cfa + loc->values[0], reg_ptr, sizeof(AddressType))) {
|
|
last_error_.code = DWARF_ERROR_MEMORY_INVALID;
|
|
last_error_.address = eval_info->cfa + loc->values[0];
|
|
return false;
|
|
}
|
|
break;
|
|
case DWARF_LOCATION_VAL_OFFSET:
|
|
*reg_ptr = eval_info->cfa + loc->values[0];
|
|
break;
|
|
case DWARF_LOCATION_REGISTER: {
|
|
uint32_t cur_reg = loc->values[0];
|
|
if (cur_reg >= eval_info->regs_info.Total()) {
|
|
last_error_.code = DWARF_ERROR_ILLEGAL_VALUE;
|
|
return false;
|
|
}
|
|
*reg_ptr = eval_info->regs_info.Get(cur_reg) + loc->values[1];
|
|
break;
|
|
}
|
|
case DWARF_LOCATION_EXPRESSION:
|
|
case DWARF_LOCATION_VAL_EXPRESSION: {
|
|
AddressType value;
|
|
bool is_dex_pc = false;
|
|
if (!EvalExpression(*loc, regular_memory, &value, &eval_info->regs_info, &is_dex_pc)) {
|
|
return false;
|
|
}
|
|
if (loc->type == DWARF_LOCATION_EXPRESSION) {
|
|
if (!regular_memory->ReadFully(value, reg_ptr, sizeof(AddressType))) {
|
|
last_error_.code = DWARF_ERROR_MEMORY_INVALID;
|
|
last_error_.address = value;
|
|
return false;
|
|
}
|
|
} else {
|
|
*reg_ptr = value;
|
|
if (is_dex_pc) {
|
|
eval_info->regs_info.regs->set_dex_pc(value);
|
|
}
|
|
}
|
|
break;
|
|
}
|
|
case DWARF_LOCATION_UNDEFINED:
|
|
if (reg == eval_info->cie->return_address_register) {
|
|
eval_info->return_address_undefined = true;
|
|
}
|
|
break;
|
|
case DWARF_LOCATION_PSEUDO_REGISTER:
|
|
last_error_.code = DWARF_ERROR_ILLEGAL_VALUE;
|
|
return false;
|
|
default:
|
|
break;
|
|
}
|
|
|
|
return true;
|
|
}
|
|
|
|
template <typename AddressType>
|
|
bool DwarfSectionImpl<AddressType>::Eval(const DwarfCie* cie, Memory* regular_memory,
|
|
const dwarf_loc_regs_t& loc_regs, Regs* regs,
|
|
bool* finished) {
|
|
RegsImpl<AddressType>* cur_regs = reinterpret_cast<RegsImpl<AddressType>*>(regs);
|
|
if (cie->return_address_register >= cur_regs->total_regs()) {
|
|
last_error_.code = DWARF_ERROR_ILLEGAL_VALUE;
|
|
return false;
|
|
}
|
|
|
|
// Get the cfa value;
|
|
auto cfa_entry = loc_regs.find(CFA_REG);
|
|
if (cfa_entry == loc_regs.end()) {
|
|
last_error_.code = DWARF_ERROR_CFA_NOT_DEFINED;
|
|
return false;
|
|
}
|
|
|
|
// Always set the dex pc to zero when evaluating.
|
|
cur_regs->set_dex_pc(0);
|
|
|
|
// Reset necessary pseudo registers before evaluation.
|
|
// This is needed for ARM64, for example.
|
|
regs->ResetPseudoRegisters();
|
|
|
|
EvalInfo<AddressType> eval_info{.loc_regs = &loc_regs,
|
|
.cie = cie,
|
|
.regular_memory = regular_memory,
|
|
.regs_info = RegsInfo<AddressType>(cur_regs)};
|
|
const DwarfLocation* loc = &cfa_entry->second;
|
|
// Only a few location types are valid for the cfa.
|
|
switch (loc->type) {
|
|
case DWARF_LOCATION_REGISTER:
|
|
if (loc->values[0] >= cur_regs->total_regs()) {
|
|
last_error_.code = DWARF_ERROR_ILLEGAL_VALUE;
|
|
return false;
|
|
}
|
|
eval_info.cfa = (*cur_regs)[loc->values[0]];
|
|
eval_info.cfa += loc->values[1];
|
|
break;
|
|
case DWARF_LOCATION_VAL_EXPRESSION: {
|
|
AddressType value;
|
|
if (!EvalExpression(*loc, regular_memory, &value, &eval_info.regs_info, nullptr)) {
|
|
return false;
|
|
}
|
|
// There is only one type of valid expression for CFA evaluation.
|
|
eval_info.cfa = value;
|
|
break;
|
|
}
|
|
default:
|
|
last_error_.code = DWARF_ERROR_ILLEGAL_VALUE;
|
|
return false;
|
|
}
|
|
|
|
for (const auto& entry : loc_regs) {
|
|
uint32_t reg = entry.first;
|
|
// Already handled the CFA register.
|
|
if (reg == CFA_REG) continue;
|
|
|
|
AddressType* reg_ptr;
|
|
if (reg >= cur_regs->total_regs()) {
|
|
if (entry.second.type != DWARF_LOCATION_PSEUDO_REGISTER) {
|
|
// Skip this unknown register.
|
|
continue;
|
|
}
|
|
if (!eval_info.regs_info.regs->SetPseudoRegister(reg, entry.second.values[0])) {
|
|
last_error_.code = DWARF_ERROR_ILLEGAL_VALUE;
|
|
return false;
|
|
}
|
|
} else {
|
|
reg_ptr = eval_info.regs_info.Save(reg);
|
|
if (!EvalRegister(&entry.second, reg, reg_ptr, &eval_info)) {
|
|
return false;
|
|
}
|
|
}
|
|
}
|
|
|
|
// Find the return address location.
|
|
if (eval_info.return_address_undefined) {
|
|
cur_regs->set_pc(0);
|
|
} else {
|
|
cur_regs->set_pc((*cur_regs)[cie->return_address_register]);
|
|
}
|
|
|
|
// If the pc was set to zero, consider this the final frame. Exception: if
|
|
// this is the sigreturn frame, then we want to try to recover the real PC
|
|
// using the return address (from LR or the stack), so keep going.
|
|
*finished = (cur_regs->pc() == 0 && !cie->is_signal_frame) ? true : false;
|
|
|
|
cur_regs->set_sp(eval_info.cfa);
|
|
|
|
return true;
|
|
}
|
|
|
|
template <typename AddressType>
|
|
bool DwarfSectionImpl<AddressType>::GetCfaLocationInfo(uint64_t pc, const DwarfFde* fde,
|
|
dwarf_loc_regs_t* loc_regs, ArchEnum arch) {
|
|
DwarfCfa<AddressType> cfa(&memory_, fde, arch);
|
|
|
|
// Look for the cached copy of the cie data.
|
|
auto reg_entry = cie_loc_regs_.find(fde->cie_offset);
|
|
if (reg_entry == cie_loc_regs_.end()) {
|
|
if (!cfa.GetLocationInfo(pc, fde->cie->cfa_instructions_offset, fde->cie->cfa_instructions_end,
|
|
loc_regs)) {
|
|
last_error_ = cfa.last_error();
|
|
return false;
|
|
}
|
|
cie_loc_regs_[fde->cie_offset] = *loc_regs;
|
|
}
|
|
cfa.set_cie_loc_regs(&cie_loc_regs_[fde->cie_offset]);
|
|
if (!cfa.GetLocationInfo(pc, fde->cfa_instructions_offset, fde->cfa_instructions_end, loc_regs)) {
|
|
last_error_ = cfa.last_error();
|
|
return false;
|
|
}
|
|
return true;
|
|
}
|
|
|
|
template <typename AddressType>
|
|
bool DwarfSectionImpl<AddressType>::Log(uint8_t indent, uint64_t pc, const DwarfFde* fde,
|
|
ArchEnum arch) {
|
|
DwarfCfa<AddressType> cfa(&memory_, fde, arch);
|
|
|
|
// Always print the cie information.
|
|
const DwarfCie* cie = fde->cie;
|
|
if (!cfa.Log(indent, pc, cie->cfa_instructions_offset, cie->cfa_instructions_end)) {
|
|
last_error_ = cfa.last_error();
|
|
return false;
|
|
}
|
|
if (!cfa.Log(indent, pc, fde->cfa_instructions_offset, fde->cfa_instructions_end)) {
|
|
last_error_ = cfa.last_error();
|
|
return false;
|
|
}
|
|
return true;
|
|
}
|
|
|
|
template <typename AddressType>
|
|
bool DwarfSectionImpl<AddressType>::Init(uint64_t offset, uint64_t size, int64_t section_bias) {
|
|
section_bias_ = section_bias;
|
|
entries_offset_ = offset;
|
|
next_entries_offset_ = offset;
|
|
entries_end_ = offset + size;
|
|
|
|
memory_.clear_func_offset();
|
|
memory_.clear_text_offset();
|
|
memory_.set_cur_offset(offset);
|
|
pc_offset_ = offset;
|
|
|
|
return true;
|
|
}
|
|
|
|
// Create a cached version of the fde information such that it is a std::map
|
|
// that is indexed by end pc and contains a pair that represents the start pc
|
|
// followed by the fde object. The fde pointers are owned by fde_entries_
|
|
// and not by the map object.
|
|
// It is possible for an fde to be represented by multiple entries in
|
|
// the map. This can happen if the the start pc and end pc overlap already
|
|
// existing entries. For example, if there is already an entry of 0x400, 0x200,
|
|
// and an fde has a start pc of 0x100 and end pc of 0x500, two new entries
|
|
// will be added: 0x200, 0x100 and 0x500, 0x400.
|
|
template <typename AddressType>
|
|
void DwarfSectionImpl<AddressType>::InsertFde(const DwarfFde* fde) {
|
|
uint64_t start = fde->pc_start;
|
|
uint64_t end = fde->pc_end;
|
|
auto it = fdes_.upper_bound(start);
|
|
while (it != fdes_.end() && start < end && it->second.first < end) {
|
|
if (start < it->second.first) {
|
|
fdes_[it->second.first] = std::make_pair(start, fde);
|
|
}
|
|
start = it->first;
|
|
++it;
|
|
}
|
|
if (start < end) {
|
|
fdes_[end] = std::make_pair(start, fde);
|
|
}
|
|
}
|
|
|
|
template <typename AddressType>
|
|
bool DwarfSectionImpl<AddressType>::GetNextCieOrFde(const DwarfFde** fde_entry) {
|
|
uint64_t start_offset = next_entries_offset_;
|
|
|
|
memory_.set_data_offset(entries_offset_);
|
|
memory_.set_cur_offset(next_entries_offset_);
|
|
uint32_t value32;
|
|
if (!memory_.ReadBytes(&value32, sizeof(value32))) {
|
|
last_error_.code = DWARF_ERROR_MEMORY_INVALID;
|
|
last_error_.address = memory_.cur_offset();
|
|
return false;
|
|
}
|
|
|
|
uint64_t cie_offset;
|
|
uint8_t cie_fde_encoding;
|
|
bool entry_is_cie = false;
|
|
if (value32 == static_cast<uint32_t>(-1)) {
|
|
// 64 bit entry.
|
|
uint64_t value64;
|
|
if (!memory_.ReadBytes(&value64, sizeof(value64))) {
|
|
last_error_.code = DWARF_ERROR_MEMORY_INVALID;
|
|
last_error_.address = memory_.cur_offset();
|
|
return false;
|
|
}
|
|
|
|
next_entries_offset_ = memory_.cur_offset() + value64;
|
|
// Read the Cie Id of a Cie or the pointer of the Fde.
|
|
if (!memory_.ReadBytes(&value64, sizeof(value64))) {
|
|
last_error_.code = DWARF_ERROR_MEMORY_INVALID;
|
|
last_error_.address = memory_.cur_offset();
|
|
return false;
|
|
}
|
|
|
|
if (value64 == cie64_value_) {
|
|
entry_is_cie = true;
|
|
cie_fde_encoding = DW_EH_PE_sdata8;
|
|
} else {
|
|
cie_offset = GetCieOffsetFromFde64(value64);
|
|
}
|
|
} else {
|
|
next_entries_offset_ = memory_.cur_offset() + value32;
|
|
|
|
// 32 bit Cie
|
|
if (!memory_.ReadBytes(&value32, sizeof(value32))) {
|
|
last_error_.code = DWARF_ERROR_MEMORY_INVALID;
|
|
last_error_.address = memory_.cur_offset();
|
|
return false;
|
|
}
|
|
|
|
if (value32 == cie32_value_) {
|
|
entry_is_cie = true;
|
|
cie_fde_encoding = DW_EH_PE_sdata4;
|
|
} else {
|
|
cie_offset = GetCieOffsetFromFde32(value32);
|
|
}
|
|
}
|
|
|
|
if (entry_is_cie) {
|
|
auto entry = cie_entries_.find(start_offset);
|
|
if (entry == cie_entries_.end()) {
|
|
DwarfCie* cie = &cie_entries_[start_offset];
|
|
cie->lsda_encoding = DW_EH_PE_omit;
|
|
cie->cfa_instructions_end = next_entries_offset_;
|
|
cie->fde_address_encoding = cie_fde_encoding;
|
|
|
|
if (!FillInCie(cie)) {
|
|
cie_entries_.erase(start_offset);
|
|
return false;
|
|
}
|
|
}
|
|
*fde_entry = nullptr;
|
|
} else {
|
|
auto entry = fde_entries_.find(start_offset);
|
|
if (entry != fde_entries_.end()) {
|
|
*fde_entry = &entry->second;
|
|
} else {
|
|
DwarfFde* fde = &fde_entries_[start_offset];
|
|
fde->cfa_instructions_end = next_entries_offset_;
|
|
fde->cie_offset = cie_offset;
|
|
|
|
if (!FillInFde(fde)) {
|
|
fde_entries_.erase(start_offset);
|
|
return false;
|
|
}
|
|
*fde_entry = fde;
|
|
}
|
|
}
|
|
return true;
|
|
}
|
|
|
|
template <typename AddressType>
|
|
void DwarfSectionImpl<AddressType>::GetFdes(std::vector<const DwarfFde*>* fdes) {
|
|
// Loop through the already cached entries.
|
|
uint64_t entry_offset = entries_offset_;
|
|
while (entry_offset < next_entries_offset_) {
|
|
auto cie_it = cie_entries_.find(entry_offset);
|
|
if (cie_it != cie_entries_.end()) {
|
|
entry_offset = cie_it->second.cfa_instructions_end;
|
|
} else {
|
|
auto fde_it = fde_entries_.find(entry_offset);
|
|
if (fde_it == fde_entries_.end()) {
|
|
// No fde or cie at this entry, should not be possible.
|
|
return;
|
|
}
|
|
entry_offset = fde_it->second.cfa_instructions_end;
|
|
fdes->push_back(&fde_it->second);
|
|
}
|
|
}
|
|
|
|
while (next_entries_offset_ < entries_end_) {
|
|
const DwarfFde* fde;
|
|
if (!GetNextCieOrFde(&fde)) {
|
|
break;
|
|
}
|
|
if (fde != nullptr) {
|
|
InsertFde(fde);
|
|
fdes->push_back(fde);
|
|
}
|
|
|
|
if (next_entries_offset_ < memory_.cur_offset()) {
|
|
// Simply consider the processing done in this case.
|
|
break;
|
|
}
|
|
}
|
|
}
|
|
|
|
template <typename AddressType>
|
|
const DwarfFde* DwarfSectionImpl<AddressType>::GetFdeFromPc(uint64_t pc) {
|
|
// Search in the list of fdes we already have.
|
|
auto it = fdes_.upper_bound(pc);
|
|
if (it != fdes_.end()) {
|
|
if (pc >= it->second.first) {
|
|
return it->second.second;
|
|
}
|
|
}
|
|
|
|
// The section might have overlapping pcs in fdes, so it is necessary
|
|
// to do a linear search of the fdes by pc. As fdes are read, a cached
|
|
// search map is created.
|
|
while (next_entries_offset_ < entries_end_) {
|
|
const DwarfFde* fde;
|
|
if (!GetNextCieOrFde(&fde)) {
|
|
return nullptr;
|
|
}
|
|
if (fde != nullptr) {
|
|
InsertFde(fde);
|
|
if (pc >= fde->pc_start && pc < fde->pc_end) {
|
|
return fde;
|
|
}
|
|
}
|
|
|
|
if (next_entries_offset_ < memory_.cur_offset()) {
|
|
// Simply consider the processing done in this case.
|
|
break;
|
|
}
|
|
}
|
|
return nullptr;
|
|
}
|
|
|
|
// Explicitly instantiate DwarfSectionImpl
|
|
template class DwarfSectionImpl<uint32_t>;
|
|
template class DwarfSectionImpl<uint64_t>;
|
|
|
|
// Explicitly instantiate DwarfDebugFrame
|
|
template class DwarfDebugFrame<uint32_t>;
|
|
template class DwarfDebugFrame<uint64_t>;
|
|
|
|
// Explicitly instantiate DwarfEhFrame
|
|
template class DwarfEhFrame<uint32_t>;
|
|
template class DwarfEhFrame<uint64_t>;
|
|
|
|
} // namespace unwindstack
|