objcopy Command in Linux
The objcopy command in Linux copies and translates the object files. It is primarily used to copy or modify object files in various ways, such as changing formats, stripping symbols, removing sections, or adding new sections. The objcopy command uses the GNU BFD Library to read and write object files and can handle most formats automatically.
Table of Contents
Here is a comprehensive guide to the options available with the objcopy command −
- Installation of objcopy Command
- Syntax of objcopy Command
- objcopy Command Options
- Examples of objcopy Command in Linux
Note − The behavior of the objcopy command may vary slightly across systems due to the handling of architecture-specific formats.
Installation of objcopy Command
By default, the objcopy tool may not be available in Linux to install it, use the instructions given below −
To install objcopy on Ubuntu, Kali Linux, Debian, and other Debian-based distributions, use the following command −
sudo apt install binutils
To install it on Arch Linux, use the command given below −
sudo pacman -S aarch64-linux-gnu-binutils
To install it on CentOS, use −
sudo yum install binutils
To install it on Fedora, use the following command −
sudo dnf install binutils-arc-linux-gnu
To verify the installation, check the version of the objcopy command −
objcopy -V
Or check the binary path using the which command −
which objcopy
Syntax of objcopy Command
The syntax of the objcopy command is as follows −
objcopy [options] [input-file] [output-file]
The [options] field is used to specify various options to change the command's behavior. The [input-file] is the object file, and the [output-file] is the file in which the modified content of the input file will be saved.
objcopy Command Options
The options of the Linux objcopy command are listed below −
| Flags | Options | Description |
|---|---|---|
| -I | --input-target <bfdname> | Specify the input file format as <bfdname> |
| -O | --output-target <bfdname> | Specify the output file format as <bfdname> |
| -B | --binary-architecture <arch> | Set the output architecture if the input lacks one |
| -F | --target <bfdname> | Set both input and output formats to <bfdname> |
| --debugging | Convert debugging information if possible | |
| --add-gnu-debuglink=<file> | Add .gnu_debuglink section linking to <file> | |
| -D | --enable-deterministic-archives | Produce deterministic archives (default) |
| -U | --disable-deterministic-archives | Disable deterministic archive behavior |
| -j | --only-section <name> | Copy only section <name> to the output |
| -K | --keep-symbol <name> | Retain symbol <name> without stripping |
| --keep-section-symbols | Retain all section symbols | |
| --keep-file-symbols | Retain all file symbols | |
| --localize-hidden | Convert all ELF hidden symbols to local symbols | |
| -L | --localize-symbol <name> | Explicitly mark symbol <name> as local |
| --globalize-symbol <name> | Explicitly mark symbol <name> as global | |
| -G | --keep-global-symbol <name> | Localize all symbols except <name> |
| -p | --preserve-dates | Copy modified/access timestamps to the output |
| -R | --remove-section <name> | Remove section <name> from the output |
| --remove-relocations <name> | Remove relocations from section <name> | |
| --strip-section-headers | Strip section headers | |
| --strip-dwo | Remove all .dwo sections | |
| --extract-dwo | Copy only .dwo sections | |
| -S | --strip-all | Remove all symbols and relocation information |
| -g | --strip-debug | Remove all debugging symbols and sections |
| --strip-unneeded | Remove symbols not needed for relocations | |
| -N | --strip-symbol <name> | Exclude symbol <name> |
| --only-keep-debug | Retain only debug information | |
| --redefine-sym <old>=<new> | Rename symbol <old> to <new> | |
| --pad-to <addr> | Pad the last section up to <addr> | |
| --set-start <addr> | Set the start address to <addr> | |
| --rename-section old=new | Renames a section | |
| -v | --verbose | Display details of modified files |
| -V | --version | To display the command's version |
| --decompress-debug-sections | Decompress DWARF debug sections | |
| -W | --weaken-symbol <name> | Mark symbol <name> as weak |
| --weaken | Mark all global symbols as weak | |
| -w | --wildcard | Allow wildcard usage in symbol-matching |
| -x | --discard-all | Remove all non-global symbols |
| -X | --discard-locals | Remove all compiler-generated symbols |
| -i | --interleave[=<number>] | Copy only N out of every <number> bytes |
| --interleave-width <number> | Define N for use with the --interleave option | |
| -h | --help | Show help information |
| @<file> | Read options from <file> | |
| --info | List supported object formats and architectures |
Examples of objcopy Command in Linux
This section demonstrates the usage of the objcopy command in Linux with examples −
Converting a File to another Format
To convert a file to another format, use the -O or --output-target option with the objcopy command −
objcopy -O binary code output.bin
The above command converts the code which is an elf to binary. Other formats that can be used to convert an elf file are listed below −
| Format | Description |
|---|---|
| elf32-mips | ELF file for the 32-bit MIPS architecture |
| elf64-mips | ELF file for 64-bit MIPS architecture |
| srec | Motorola S-records format |
| ihex | Intel Hexadecimal object file format |
| binary | Raw binary format (no headers or metadata) |
| tekhex | Tektronix hexadecimal object file format |
| a.out-i386 | a.out file format for 32-bit Intel x86 |
| elf64-x86-64 | ELF file for 64-bit x86-64 architecture |
| pei-i386 | PE/COFF file for 32-bit Intel x86 (Windows executables) |
| pei-x86-64 | PE/COFF file for 64-bit x86-64 |
| elf32-arm | ELF file for 32-bit ARM architecture |
These formats are dependent on the architecture of the system.
Stripping Debug Symbol
To strip the debug symbol, use the --strip-debug option with the objcopy command −
objcopy --strip-debug code output.elf
In the above command, the code is the input file, while output.elf is the output file with stripped debug information.
Extracting a Specific Section
To extract a specific section from the object file, use the -j or --only-section option. For example, to extract .text section use the following command −
objcopy -j .text -O binary code text_section.bin
Removing a Specific Section
To remove a specific section, from the object file, use the -R or --remove-section option. For example, to remove the .data section use the command given below −
objcopy -R .data code output.elf
Saving Debug Information in a Separate File
To keep the debug information in a separate file, use the command given below −
objcopy --only-keep-debug code output.elf
Removing Unused Symbols
To remove all the unused symbols, use the --strip-unneeded option −
objcopy --strip-unneeded code output.elf
Renaming a Section
To rename a section, use the --rename-section option. For example, to rename the .text section to .mytext use the following command −
objcopy --rename-section .text=.mytext code output.elf
Preserving the Dates
To preserve the access and modification timestamps of the output file, matching them to those of the input file, use the -p or --preserve-dates option −
objcopy -p code output.elf
Marking a Symbol as Weak
To mark a symbol as weak, use the --weaken-symbol option with the objcopy command −
objcopy --weaken-symbol=symbol code output.elf
The symbol in the above command can be a function or variable.
Displaying Help
To display help related to the objcopy command, use the -h or --help option −
objcopy --help
Conclusion
The objcopy command in Linux is a handy tool for manipulating object files. It supports various options for customizing the behavior, including the ability to extract or remove specific sections, rename them, and manage debugging information.
Overall, objcopy is an essential utility for managing and optimizing object files in Linux environments.