bitfields

Bitfields are a feature of certain programming languages, notably C and C++, that allow the declaration of structure members with a specific number of bits. A bit-field is declared by giving a width after a colon, such as unsigned int flag:1;. Bitfields enable compact storage of small integers and boolean flags by packing them into a single storage unit, typically a byte or word. They are commonly used in hardware register representations and in network protocol headers where precise bit-level control is required.

Key properties and limitations: The type for a bit-field must be an integral type. The width can be between 1 and the width of the underlying type. The allocation, ordering, and exact storage size of bit-fields are implementation-defined. Endianness and bit ordering are not standardized across platforms, and compilers may place bit-fields in different ways. Some languages or compilers do not permit taking the address of a bit-field. The total width of a set of bit-fields within a single storage unit cannot be assumed portable; one should rely on the compiler’s documented behavior.

Usage and alternatives: Bitfields provide convenient compact representations but can complicate portability and alignment. They are not guaranteed to be thread-safe or atomic. For portable bit-level manipulation across platforms, explicit masks and shifts on unsigned integers are often preferred. Example in C: struct Status { unsigned int ready:1; unsigned int error:1; unsigned int mode:2; unsigned int reserved:28; };

Limitations: Some compilers may insert padding between fields or between fields and the containing type, and the order of allocation is implementation-defined. They are best used when the exact layout is known for a single toolchain or when interfacing with hardware where the layout is fixed and documented.