1Large data sections 2=================== 3 4When linking very large binaries, lld may report relocation overflows like 5 6:: 7 8 relocation R_X86_64_PC32 out of range: 2158227201 is not in [-2147483648, 2147483647] 9 10This happens when running into architectural limitations. For example, in x86-64 11PIC code, a reference to a static global variable is typically done with a 12``R_X86_64_PC32`` relocation, which is a 32-bit signed offset from the PC. That 13means if the global variable is laid out further than 2GB (2^31 bytes) from the 14instruction referencing it, we run into a relocation overflow. 15 16lld normally lays out sections as follows: 17 18.. image:: section_layout.png 19 20The largest relocation pressure is usually from ``.text`` to the beginning of 21``.rodata`` or ``.text`` to the end of ``.bss``. 22 23Some code models offer a tradeoff between relocation pressure and performance. 24For example, x86-64's medium code model splits global variables into small and 25large globals depending on if their size is over a certain threshold. Large 26globals are placed further away from text and we use 64-bit references to refer 27to them. 28 29Large globals are placed in separate sections from small globals, and those 30sections have a "large" section flag, e.g. ``SHF_X86_64_LARGE`` for x86-64. The 31linker places large sections on the outer edges of the binary, making sure they 32do not affect affect the distance of small globals to text. The large versions 33of ``.rodata``, ``.bss``, and ``.data`` are ``.lrodata``, ``.lbss``, and 34``.ldata``, and they are laid out as follows: 35 36.. image:: large_section_layout_pic.png 37 38We try to keep the number of ``PT_LOAD`` segments to a minimum, so we place 39large sections next to the small sections with the same RWX permissions when 40possible. 41 42``.lbss`` is right after ``.bss`` so that they are merged together and we 43minimize the number of segments with ``p_memsz > p_filesz``. 44 45Note that the above applies to PIC code. For less common non-PIC code with 46absolute relocations instead of relative relocations, 32-bit relocations 47typically assume that symbols are in the lower 2GB of the address space. So for 48non-PIC code, large sections should be placed after all small sections to avoid 49``.lrodata`` pushing small symbols out of the lower 2GB of the address space. 50``-z lrodata-after-bss`` changes the layout to be: 51 52.. image:: large_section_layout_nopic.png 53
