diff options
Diffstat (limited to 'src/endian.h')
| -rw-r--r-- | src/endian.h | 74 |
1 files changed, 63 insertions, 11 deletions
diff --git a/src/endian.h b/src/endian.h index 11712b5..0b3e8af 100644 --- a/src/endian.h +++ b/src/endian.h @@ -1,3 +1,14 @@ +/** \file + * \brief Proone's own <endian.h>. Proone does not use <endian.h> present on + * many platforms simply because the header is not a standard. Proone does not + * use the traditional byte swap approach to solve the endianness problem for + * data communication so that all arches have the same disadvantage when + * communicating with other hosts. + * \note The functions in the internet protocol headers are used if present to + * avoid confusion. + * \see BYTEORDER(3) + * \see <arpa/inet.h> + */ /* * Copyright (c) 2019-2021 David Timber <mieabby@gmail.com> * @@ -20,23 +31,43 @@ * SOFTWARE. */ #pragma once -/********************************************************************** -* Endianess Independent Byte Extraction -***********************************************************************/ -/* prne_getmsbN(x, n) -* -* Extract nth most significant byte of x. -*/ + +/** \def prne_getmsb \def prne_getmsb64 \def prne_getmsb32 \def prne_getmsb16 + * \brief Extract \p n th most significant byte of \p x + * \param[in] x The integer + * \param[in] n The byte place, in range of [0, 1] if \p x is 16-bit integer, + * [0, 3] if \p x is 32-bit integer, [0, 7] if \p x is 64-bit integer. + * \param[in] w The data type to use in calculation. One of uint_fastN_t + * variants. + * \param[in] s The number of bits to shift by. + * \return The 8-bit integer extracted from the \p n th place of \p x in range + * [0, 255]. + * \see \c prne_recmb_msb64() + * \see \c prne_recmb_msb32() + * \see \c prne_recmb_msb16() + */ #define prne_getmsb(x, n, w, s)\ (uint8_t)(((w)(x) & (w)0xFF << (s - 8 * (n))) >> (s - 8 * (n))) #define prne_getmsb64(x, n) prne_getmsb((x), (n), uint_fast64_t, 56) #define prne_getmsb32(x, n) prne_getmsb((x), (n), uint_fast32_t, 24) #define prne_getmsb16(x, n) prne_getmsb((x), (n), uint_fast16_t, 8) -/* prne_recmb_msbN(...) -* -* Recombine bytes in big-endian order to uintN. -*/ +/** \def prne_recmb_msb64 \def prne_recmb_msb32 \def prne_recmb_msb16 + * \brief Recombine bytes in big-endian order. + * \param a The first byte. + * \param b The second byte. + * \param c The third byte. + * \param d The fourth byte. + * \param e The fifth byte. + * \param f The sixth byte. + * \param g The seventh byte. + * \param h The eighth byte. + * \return The recombined integer in the host's endian. + * \see \c prne_getmsb() + * \see \c prne_getmsb64() + * \see \c prne_getmsb32() + * \see \c prne_getmsb16() + */ #define prne_recmb_msb64(a, b, c, d, e, f, g, h) (\ ((uint_fast64_t)(a) << 56) |\ ((uint_fast64_t)(b) << 48) |\ @@ -60,9 +91,18 @@ /* Machine Characteristics */ +/** \def PRNE_ENDIAN_LITTLE \def PRNE_ENDIAN_BIG + * \brief The values that can be matched against \c PRNE_HOST_ENDIAN + * \note The PDP endian is not defined because the ELF does not support it. + */ #define PRNE_ENDIAN_LITTLE 1 #define PRNE_ENDIAN_BIG 2 +/** \def PRNE_HOST_ENDIAN + * \brief The host endian. + * \see \c PRNE_ENDIAN_LITTLE + * \see \c PRNE_ENDIAN_BIG + */ #ifdef __GNUC__ #if __BYTE_ORDER__ == __ORDER_BIG_ENDIAN__ #define PRNE_HOST_ENDIAN PRNE_ENDIAN_BIG @@ -72,11 +112,23 @@ #error "FIXME!" #endif #else + // Expand this to support compilers other than GCC #error "FIXME!" #endif +/** + * \brief Swap bytes to invert the endianness of the 16-bit integer. + * \param x The integer. + * \returns The integer with its bytes swapped. + */ #define prne_einv16(x) (((0xFF00 & x) >> 8) | ((0x00FF & x) << 8)) +/** \def prne_htobe16 \def prne_be16toh \def prne_htole16 \def prne_le16toh + * \brief Convert the endianness of the integer. + * \param x The integer. + * \return The integer converted. + * \note Use the functions in <arpa/inet.h> where appropriate! + */ #if PRNE_HOST_ENDIAN == PRNE_ENDIAN_BIG #define prne_htobe16(x) (x) #define prne_be16toh(x) (x) |