/* * 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 #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include "DwarfCfa.h" #include "DwarfDebugFrame.h" #include "DwarfEhFrame.h" #include "DwarfEncoding.h" #include "DwarfOp.h" #include "RegsInfo.h" namespace unwindstack { DwarfSection::DwarfSection(std::shared_ptr& 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. DwarfLocations 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 const DwarfCie* DwarfSectionImpl::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 bool DwarfSectionImpl::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(-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; // TODO(b/192012848): This is wrong. We need to propagate pointer size here. cie->fde_address_encoding = DW_EH_PE_udata8; 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; // TODO(b/192012848): This is wrong. We need to propagate pointer size here. cie->fde_address_encoding = DW_EH_PE_udata4; 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 bool DwarfSectionImpl::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) { char address_size; if (!memory_.ReadBytes(&address_size, 1)) { last_error_.code = DWARF_ERROR_MEMORY_INVALID; last_error_.address = memory_.cur_offset(); return false; } cie->fde_address_encoding = address_size == 8 ? DW_EH_PE_udata8 : DW_EH_PE_udata4; // 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(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 const DwarfFde* DwarfSectionImpl::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 bool DwarfSectionImpl::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(-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 bool DwarfSectionImpl::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(cie->fde_address_encoding, &fde->pc_start); fde->pc_start = AdjustPcFromFde(fde->pc_start); memory_.set_pc_offset(0); if (!valid || !memory_.ReadEncodedValue(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(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 bool DwarfSectionImpl::EvalExpression(const DwarfLocation& loc, Memory* regular_memory, AddressType* value, RegsInfo* regs_info, bool* is_dex_pc) { DwarfOp 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 struct EvalInfo { const DwarfLocations* loc_regs; const DwarfCie* cie; Memory* regular_memory; AddressType cfa; bool return_address_undefined = false; RegsInfo regs_info; }; template bool DwarfSectionImpl::EvalRegister(const DwarfLocation* loc, uint32_t reg, AddressType* reg_ptr, void* info) { EvalInfo* eval_info = reinterpret_cast*>(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 bool DwarfSectionImpl::Eval(const DwarfCie* cie, Memory* regular_memory, const DwarfLocations& loc_regs, Regs* regs, bool* finished) { RegsImpl* cur_regs = reinterpret_cast*>(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 eval_info{.loc_regs = &loc_regs, .cie = cie, .regular_memory = regular_memory, .regs_info = RegsInfo(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 bool DwarfSectionImpl::GetCfaLocationInfo(uint64_t pc, const DwarfFde* fde, DwarfLocations* loc_regs, ArchEnum arch) { DwarfCfa 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 bool DwarfSectionImpl::Log(uint8_t indent, uint64_t pc, const DwarfFde* fde, ArchEnum arch) { DwarfCfa 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 bool DwarfSectionImpl::Init(const SectionInfo& info) { if (info.flags & SHF_COMPRESSED) { return false; } section_bias_ = info.bias; entries_offset_ = info.offset; entries_end_ = info.offset + info.size; memory_.clear_func_offset(); memory_.clear_text_offset(); memory_.set_cur_offset(info.offset); pc_offset_ = info.offset; return true; } // Read CIE or FDE entry at the given offset, and set the offset to the following entry. // The 'fde' argument is set only if we have seen an FDE entry. template bool DwarfSectionImpl::GetNextCieOrFde(uint64_t& next_entries_offset, std::optional& fde_entry) { const 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(-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_udata8; } 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_udata4; } 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.reset(); } else { fde_entry = DwarfFde{}; fde_entry->cfa_instructions_end = next_entries_offset; fde_entry->cie_offset = cie_offset; if (!FillInFde(&*fde_entry)) { return false; } } return true; } template void DwarfSectionImpl::GetFdes(std::vector* fdes) { if (fde_index_.empty()) { BuildFdeIndex(); } for (auto& it : fde_index_) { fdes->push_back(GetFdeFromOffset(it.second)); } } template const DwarfFde* DwarfSectionImpl::GetFdeFromPc(uint64_t pc) { // Ensure that the binary search table is initialized. if (fde_index_.empty()) { BuildFdeIndex(); } // Find the FDE offset in the binary search table. auto comp = [](uint64_t pc, auto& entry) { return pc < entry.first; }; auto it = std::upper_bound(fde_index_.begin(), fde_index_.end(), pc, comp); if (it == fde_index_.end()) { return nullptr; } // Load the full FDE entry based on the offset. const DwarfFde* fde = GetFdeFromOffset(/*fde_offset=*/it->second); return fde != nullptr && fde->pc_start <= pc ? fde : nullptr; } // Create binary search table to make FDE lookups fast (sorted by pc_end). // We store only the FDE offset rather than the full entry to save memory. // // If there are overlapping entries, it inserts additional entries to ensure // that one of the overlapping entries is found (it is undefined which one). template void DwarfSectionImpl::BuildFdeIndex() { struct FdeInfo { uint64_t pc_start, pc_end, fde_offset; }; std::vector fdes; for (uint64_t offset = entries_offset_; offset < entries_end_;) { const uint64_t initial_offset = offset; std::optional fde; if (!GetNextCieOrFde(offset, fde)) { break; } if (fde.has_value() && /* defensive check */ (fde->pc_start < fde->pc_end)) { fdes.push_back({fde->pc_start, fde->pc_end, initial_offset}); } if (offset <= initial_offset) { break; // Jump back. Simply consider the processing done in this case. } } std::sort(fdes.begin(), fdes.end(), [](const FdeInfo& a, const FdeInfo& b) { return std::tie(a.pc_end, a.fde_offset) < std::tie(b.pc_end, b.fde_offset); }); // If there are overlapping entries, ensure that we can always find one of them. // For example, for entries: [300, 350) [400, 450) [100, 550) [600, 650) // We add the following: [100, 300) [100, 400) // Which ensures that the [100, 550) entry can be found in its whole range. if (!fdes.empty()) { FdeInfo filling = fdes.back(); // Entry with the minimal pc_start seen so far. for (ssize_t i = fdes.size() - 1; i >= 0; i--) { // Iterate backwards. uint64_t prev_pc_end = (i > 0) ? fdes[i - 1].pc_end : 0; // If there is a gap between entries and the filling reaches the gap, fill it. if (prev_pc_end < fdes[i].pc_start && filling.pc_start < fdes[i].pc_start) { fdes.push_back({filling.pc_start, fdes[i].pc_start, filling.fde_offset}); } if (fdes[i].pc_start < filling.pc_start) { filling = fdes[i]; } } } // Copy data to the final binary search table (pc_end, fde_offset) and sort it. fde_index_.reserve(fdes.size()); for (const FdeInfo& it : fdes) { fde_index_.emplace_back(it.pc_end, it.fde_offset); } if (!std::is_sorted(fde_index_.begin(), fde_index_.end())) { std::sort(fde_index_.begin(), fde_index_.end()); } } // Explicitly instantiate DwarfSectionImpl template class DwarfSectionImpl; template class DwarfSectionImpl; // Explicitly instantiate DwarfDebugFrame template class DwarfDebugFrame; template class DwarfDebugFrame; // Explicitly instantiate DwarfEhFrame template class DwarfEhFrame; template class DwarfEhFrame; } // namespace unwindstack