Skip to content

Commit

Permalink
Minimize use of ELF64LEObjectFile over ELFObjectFileBase in `anno…
Browse files Browse the repository at this point in the history 
…tating_importer`. (#241)

 * Minor refactoring of usages of `ELF64LEObjectFile`s to
   `ELFObjectFileBase`s in `annotating_importer`.
 * This change is meant to help merge some functionality in
   `annotating_importer` with `extract_bbs_from_obj_lib`.
  • Loading branch information
@virajbshah
virajbshah authored Oct 28, 2024
1 parent 2cc6a8c commit 53b4e5d
Show file tree
Hide file tree
Showing 2 changed files with 81 additions and 43 deletions.
83 changes: 45 additions & 38 deletions gematria/datasets/annotating_importer.cc
View file
Original file line number Diff line number Diff line change
Expand Up @@ -101,67 +101,40 @@ absl::Status AnnotatingImporter::LoadBinary(std::string_view file_name) {
return absl::OkStatus();
}

absl::StatusOr<llvm::object::ELF64LEObjectFile *>
absl::StatusOr<llvm::object::ELFObjectFileBase *>
AnnotatingImporter::GetELFFromBinary() {
llvm::object::Binary *binary = owning_binary_.getBinary();
if (!binary->isObject()) {
return absl::InvalidArgumentError(
absl::StrFormat("The given binary (%s) is not an object.",
std::string(binary->getFileName())));
}
llvm::object::ObjectFile *object =
llvm::cast<llvm::object::ObjectFile>(binary);
if (!object) {
auto *object = llvm::cast<llvm::object::ObjectFile>(binary);
if (object == nullptr) {
return absl::InvalidArgumentError(
absl::StrFormat("Could not cast the binary (%s) to an ObjectFile.",
std::string(binary->getFileName())));
}

// Make sure the object is an ELF file.
if (!object->isELF() || !object->is64Bit() || !object->isLittleEndian()) {
if (!object->isELF()) {
return absl::InvalidArgumentError(
absl::StrFormat("The given object (%s) is not in ELF64LE format.",
absl::StrFormat("The given object (%s) is not in ELF format.",
std::string(binary->getFileName())));
}
auto *elf_object = llvm::dyn_cast<llvm::object::ELF64LEObjectFile>(object);
if (!elf_object) {
auto *elf_object = llvm::dyn_cast<llvm::object::ELFObjectFileBase>(object);
if (elf_object == nullptr) {
return absl::InvalidArgumentError(absl::StrFormat(
"Could not cast the object (%s) to an ELF64LEObjectFile.",
"Could not cast the object (%s) to an ELFObjectFileBase.",
std::string(binary->getFileName())));
}

return elf_object;
}

absl::StatusOr<llvm::object::Elf_Phdr_Impl<llvm::object::ELF64LE>>
AnnotatingImporter::GetMainProgramHeader(
const llvm::object::ELF64LEObjectFile *elf_object) {
llvm::object::Elf_Phdr_Impl<llvm::object::ELF64LE> main_header;
bool found_main_header = false;
auto program_headers = elf_object->getELFFile().program_headers();
if (llvm::Error error = program_headers.takeError()) {
return LlvmErrorToStatus(std::move(error));
}
for (const auto &program_header : *program_headers) {
if (program_header.p_type == llvm::ELF::PT_LOAD &&
program_header.p_flags & llvm::ELF::PF_R &&
program_header.p_flags & llvm::ELF::PF_X) {
if (found_main_header) {
return absl::InvalidArgumentError(
"The given object has multiple executable segments. This is "
"currently not supported.");
}
main_header = program_header;
found_main_header = true;
}
}

return main_header;
}

absl::StatusOr<std::vector<DisassembledInstruction>>
AnnotatingImporter::GetELFSlice(
const llvm::object::ELF64LEObjectFile *elf_object, uint64_t range_begin,
const llvm::object::ELFObjectFileBase *elf_object, uint64_t range_begin,
uint64_t range_end, uint64_t file_offset) {
llvm::StringRef binary_buf = elf_object->getData();

Expand Down Expand Up @@ -195,7 +168,24 @@ AnnotatingImporter::GetBlocksFromELF() {
if (llvm::Error error = bb_addr_map.takeError()) {
return LlvmErrorToStatus(std::move(error));
}
const auto main_header = GetMainProgramHeader(*elf_object);

// TODO(vbshah): Consider making it possible to use other ELFTs rather than
// only ELF64LE since only the implementation of GetMainProgramHeader differs
// between different ELFTs.
if (!(*elf_object)->is64Bit() || !(*elf_object)->isLittleEndian()) {
return absl::InvalidArgumentError(
absl::StrFormat("The given object (%s) is not in ELF64LE format.",
(*elf_object)->getFileName()));
}
auto *typed_elf_object =
llvm::dyn_cast<llvm::object::ELF64LEObjectFile>(*elf_object);
if (typed_elf_object == nullptr) {
return absl::InvalidArgumentError(absl::StrFormat(
"Could not cast the ELF object (%s) to an ELF64LEObjectFileBase.",
(*elf_object)->getFileName()));
}

const auto main_header = GetMainProgramHeader(typed_elf_object);
if (!main_header.ok()) {
return main_header.status();
}
Expand Down Expand Up @@ -296,7 +286,24 @@ AnnotatingImporter::GetLBRBlocksWithLatency() {
if (!elf_object.ok()) {
return elf_object.status();
}
const auto main_header = GetMainProgramHeader(*elf_object);

// TODO(vbshah): Consider making it possible to use other ELFTs rather than
// only ELF64LE since only the implementation of GetMainProgramHeader differs
// between different ELFTs.
if (!(*elf_object)->is64Bit() || !(*elf_object)->isLittleEndian()) {
return absl::InvalidArgumentError(
absl::StrFormat("The given object (%s) is not in ELF64LE format.",
(*elf_object)->getFileName()));
}
auto *typed_elf_object =
llvm::dyn_cast<llvm::object::ELF64LEObjectFile>(*elf_object);
if (typed_elf_object == nullptr) {
return absl::InvalidArgumentError(absl::StrFormat(
"Could not cast the ELF object (%s) to an ELF64LEObjectFileBase.",
(*elf_object)->getFileName()));
}

const auto main_header = GetMainProgramHeader(typed_elf_object);
if (!main_header.ok()) {
return main_header.status();
}
Expand Down
41 changes: 36 additions & 5 deletions gematria/datasets/annotating_importer.h
View file
Original file line number Diff line number Diff line change
Expand Up @@ -29,10 +29,13 @@
#include "gematria/datasets/bhive_importer.h"
#include "gematria/llvm/canonicalizer.h"
#include "gematria/llvm/disassembler.h"
#include "gematria/llvm/llvm_to_absl.h"
#include "gematria/proto/throughput.pb.h"
#include "llvm/BinaryFormat/ELF.h"
#include "llvm/Object/Binary.h"
#include "llvm/Object/ELFObjectFile.h"
#include "llvm/Object/ELFTypes.h"
#include "llvm/Support/Error.h"
#include "quipper/perf_data.pb.h"
#include "quipper/perf_parser.h"
#include "quipper/perf_reader.h"
Expand Down Expand Up @@ -67,16 +70,17 @@ class AnnotatingImporter {
// Returns a pointer inside the loaded binary casted down to an ELF object.
// The pointer is owned by this instance of `AnnotatingImporter` and may only
// be accessed while this is alive.
absl::StatusOr<llvm::object::ELF64LEObjectFile*> GetELFFromBinary();
absl::StatusOr<llvm::object::ELFObjectFileBase*> GetELFFromBinary();

// Returns the file offset of the passed ELF object.
absl::StatusOr<llvm::object::Elf_Phdr_Impl<llvm::object::ELF64LE>>
GetMainProgramHeader(const llvm::object::ELF64LEObjectFile* elf_object);
// Returns the program header corresponding to the main executable section.
template <class ELFT>
absl::StatusOr<llvm::object::Elf_Phdr_Impl<ELFT>> GetMainProgramHeader(
const llvm::object::ELFObjectFile<ELFT>* elf_object);

// Disassembles and returns instructions between two addresses in an ELF
// object.
absl::StatusOr<std::vector<DisassembledInstruction>> GetELFSlice(
const llvm::object::ELF64LEObjectFile* elf_object, uint64_t range_begin,
const llvm::object::ELFObjectFileBase* elf_object, uint64_t range_begin,
uint64_t range_end, uint64_t file_offset);

// Extracts basic blocks from an ELF object, and returns them as tuple
Expand Down Expand Up @@ -112,6 +116,33 @@ class AnnotatingImporter {
llvm::object::OwningBinary<llvm::object::Binary> owning_binary_;
};

template <class ELFT>
absl::StatusOr<llvm::object::Elf_Phdr_Impl<ELFT>>
AnnotatingImporter::GetMainProgramHeader(
const llvm::object::ELFObjectFile<ELFT>* elf_object) {
llvm::object::Elf_Phdr_Impl<ELFT> main_header;
bool found_main_header = false;
auto program_headers = elf_object->getELFFile().program_headers();
if (llvm::Error error = program_headers.takeError()) {
return LlvmErrorToStatus(std::move(error));
}
for (const auto& program_header : *program_headers) {
if (program_header.p_type == llvm::ELF::PT_LOAD &&
program_header.p_flags & llvm::ELF::PF_R &&
program_header.p_flags & llvm::ELF::PF_X) {
if (found_main_header) {
return absl::InvalidArgumentError(
"The given object has multiple executable segments. This is "
"currently not supported.");
}
main_header = program_header;
found_main_header = true;
}
}

return main_header;
}

} // namespace gematria

#endif // THIRD_PARTY_GEMATRIA_GEMATRIA_DATASETS_ANNOTATING_IMPORTER_H_

0 comments on commit 53b4e5d

Please sign in to comment.