Relocation (computing)

In software development, relocation is the process of assigning load addresses for position-dependent code and data of a program and adjusting the code and data to reflect the assigned addresses.

A linker usually performs relocation in conjunction with symbol resolution, the process of searching files and libraries to replace symbolic references or names of libraries with actual usable addresses in memory before running a program.

Relocation is typically done by the linker at link time, but it can also be done at load time by a relocating loader, or at run time by the running program itself.

Segmentation

Object files are typically segmented into various memory segment or section types. Example segment types include code segment (.text), initialized data segment (.data), uninitialized data segment (.bss), or others as established by the programmer, such as common segments, or named static segments.

Relocation table

The relocation table is a list of addresses created by a compiler or assembler and stored in the object or executable file. Each entry in the table references an absolute address in the object code that must be changed when the loader relocates the program so that it will refer to the correct location. Entries in the relocation table are known as fixups and are designed to support relocation of the program as a complete unit. In some cases, each fixup in the table is itself relative to a base address of zero, so the fixups themselves must be changed as the loader moves through the table.

In some architectures, a fixup that crosses certain boundaries (such as a segment boundary) or that is not aligned on a word boundary is illegal and flagged as an error by the linker.

DOS and 16-bit Windows

Far pointers (32-bit pointers with segment:offset, used to address 20-bit 640 KB memory space available to DOS programs), which point to code or data within a DOS executable (EXE), do not have absolute segments, because the actual address of code or data depends on where the program is loaded in memory and this is not known until the program is loaded.

Instead, segments are relative values in the DOS EXE file. These segments need to be corrected, when the executable has been loaded into memory. The EXE loader uses a relocation table to find the segments that need to be adjusted.

32-bit Windows

With 32-bit Windows operating systems, it is not mandatory to provide relocation tables for EXE files, since they are the first image loaded into the virtual address space and thus will be loaded at their preferred base address.

For both DLLs and for EXEs which opt into address space layout randomization (ASLR), an exploit mitigation technique introduced with Windows Vista, relocation tables once again become mandatory because of the possibility that the binary may be dynamically moved before being executed, even though they are still the first thing loaded in the virtual address space.

64-bit Windows

When running native 64-bit binaries on Windows Vista and above, ASLR is mandatory, and thus relocation sections cannot be omitted by the compiler.

Unix-like systems

The Executable and Linkable Format (ELF) executable format and shared library format used by most Unix-like systems allows several types of relocation to be defined.

Relocation procedure

The linker reads segment information and relocation tables in the object files and performs relocation by:

• merging all segments of common type into a single segment of that type
• assigning unique run time addresses to each section and each symbol, giving all code (functions) and data (global variables) unique run time addresses Template:Clarify
• referring to the relocation table to modifyTemplate:Why symbol references so that they point to the correctTemplate:Clarify run time addresses.

Example

The following example uses Donald Knuth's MIX architecture and MIXAL assembly language. The principles are the same for any architecture, though the details will change.

File:Relocation example.tif

• (A) Program SUBR is compiled to produce object file (B), shown as both machine code and assembler. The compiler may start the compiled code at an arbitrary location, often location 1 as shown. Location 13 contains the machine code for the jump instruction to statement ST in location 5.
• (C) If SUBR is later linked with other code it may be stored at a location other than 1. In this example the linker places it at location 120. The address in the jump instruction, which is now at location 133, must be relocated to point to the new location of the code for statement ST, now 125. [1 61 shown in the instruction is the MIX machine code representation of 125].
• (D) When the program is loaded into memory to run it may be loaded at some location other than the one assigned by the linker. This example shows SUBR now at location 300. The address in the jump instruction, now at 313, needs to be relocated again so that it points to the updated location of ST, 305. [4 49 is the MIX machine representation of 305].

Alternatives

Some architectures avoid relocation entirely by deferring address assignment to run time; as, for example, in stack machines with zero address arithmetic or in some segmented architectures where every compilation unit is loaded into a separate segment.

See also

• Linker (computing)
• Library (computing)
• Object file
• Prebinding
• Static library
• Self-relocation
• Rebasing
• Garbage collection
• Pointer swizzling, a lazy form of pointer modification
• Relocatable Object Module Format

Further reading

• Template:Cite magazine (3 pages) (NB. Describes a relocatable hex format by Mostek.)
• Template:Cite magazine (8 pages) (NB. Describes a relocatable hex format by TDL.)
• [1][2][3]. Originally presented at: (609 pages). (This "resize" method, named page boundary relocation, could be applied statically to a CP/M-80 disk image using Template:Ill in order to maximize the TPA for programs to run. It was also utilized dynamically by the CP/M debugger Dynamic Debugging Tool (DDT) to relocate itself into higher memory. The same approach was independently developed by Bruce H. Van Natta of IMS Associates to produce relocatable PL/M code. As paragraph boundary relocation, another variant of this method was later utilized by dynamically HMA self-relocating TSRs like KEYB, SHARE, and NLSFUNC under DR DOS 6.0 and higher. A much more sophisticated and byte-level granular method based on a somewhat similar approach was independently conceived and implemented by Matthias R. Paul and Axel C. Frinke for their dynamic dead-code elimination to dynamically minimize the runtime footprint of resident drivers and TSRs (like FreeKEYB).)
• Template:Cite magazine [4][5] (2 pages) (NB. Describes page boundary relocation and relocating assemblers.)
• [6][7] (33 pages)
• (2+xiv+270+6 pages)
• [8]
• [9][10]
• [11][12]
• [13][14]
• [15][16]
• [17][18][19]
• [20][21][22]
• [23][24]