FILEOFFSETBITS

FILEOFFSETBITS is a macro concept used in some C libraries and kernel headers to control the width of file offsets, effectively choosing between 32-bit and 64-bit offsets. In practice, the commonly used form is _FILE_OFFSET_BITS, which, when defined before including standard headers, switches the offset type used by the C library from 32-bit to 64-bit. This enables large-file support (LFS) on systems where 32-bit offsets would limit file sizes to 2 GiB.

How it works: on 32-bit systems, defining _FILE_OFFSET_BITS as 64 causes types such as off_t to become 64-bit, and it can enable 64-bit variants of file-seeking and positioning functions (for example, fseeko/ftello and lseek) to operate on larger files. If the macro is not defined or is defined as 32, the default 32-bit offsets remain in effect. On native 64-bit systems, these offsets are typically already 64-bit, so the macro may have no practical effect.

Usage notes: to enable large-file support in user-space code, place a preprocessor directive such as #define _FILE_OFFSET_BITS 64 before including any headers. Some environments also support a corresponding _FILE_OFFSET_BITS value of 32 to contract offsets back to 32 bits. Developers should be aware of ABI implications, as enabling 64-bit offsets changes the sizes of types like off_t and can affect binary compatibility with pre-LFS code.

Variants and naming: the exact macro name can vary by library or platform, with FILE_OFFSET_BITS and FILEOFFSETBITS appearing in some contexts. The underlying idea remains the same: selecting 32-bit vs 64-bit file offsets to support large files. See also Large File Support, off_t, lseek, fseeko, and ftello.