Two libraries `libbfd` and `libelf` are used to parse a binary.

[yomaytk]

Description

elfconv needs to parse the ELF binary and uses two libraries, libbfd and libelf. However, libelf is enough to parse ELF binary and it is specific to ELF format, so we should fix not to use libbfd.
For example, elfconv uses libbfd to get every section data as follows, but maybe libelf can do the same functions.

elfconv/lifter/Binary/Loader.cpp

Lines 261 to 302 in 2853198

	void ELFObject::LoadSectionsBFD() {
	asection *bfd_sec;
	for (bfd_sec = bfd_h->sections; bfd_sec; bfd_sec = bfd_sec->next) {
	ELFSection::SectionType sec_type;
	flagword bfd_flags;
	bfd_vma vma;
	bfd_size_type size;
	std::string sec_name;
	uint8_t *sec_bytes;
	// get bfd flags
	bfd_flags = bfd_section_flags(bfd_sec);
	if (bfd_flags & SEC_CODE) {
	sec_type = ELFSection::SEC_TYPE_CODE;
	} else if (bfd_flags & (SEC_DATA | SEC_ALLOC)) {
	sec_type = ELFSection::SEC_TYPE_DATA;
	} else if (bfd_flags & SEC_READONLY) {
	sec_type = ELFSection::SEC_TYPE_READONLY;
	} else {
	sec_type = ELFSection::SEC_TYPE_UNKNOWN;
	}
	// get vma, section size, section name, section contents
	vma = bfd_section_vma(bfd_sec);
	size = bfd_section_size(bfd_sec);
	sec_name = std::string(bfd_section_name(bfd_sec));
	if (sec_name.empty()) {
	sec_name = std::string("<unnamed>");
	}
	sec_bytes = reinterpret_cast<uint8_t *>(malloc(size));
	if (!sec_bytes) {
	elfconv_runtime_error("failed to allocate section bytes.\n");
	}
	if (!bfd_get_section_contents(bfd_h, bfd_sec, sec_bytes, 0, size)) {
	elfconv_runtime_error("failed to read and copy section bytes.\n");
	}
	sections.emplace_back(this, sec_type, sec_name, vma, size, sec_bytes);
	}
	}

[Fancy2209]

Could also use ELFIO to replace the 2

[yomaytk]

Thank you for the suggestion! ELFIO looks simple and provides the features we need. In any case, we would like to remove the dependency on libbfd.