view common.h @ 728:1fa3820b1a84 libavutil

ARM asm for AV_RN*() ARMv6 and later support unaligned loads and stores for single word/halfword but not double/multiple. GCC is ignorant of this and will always use bytewise accesses for unaligned data. Casting to an int32_t pointer is dangerous since a load/store double or multiple instruction might be used (this happens with some code in FFmpeg). Implementing the AV_[RW]* macros with inline asm using only supported instructions gives fast and safe unaligned accesses. ARM RVCT does the right thing with generic code. This gives an overall speedup of up to 10%.

/*
 * copyright (c) 2006 Michael Niedermayer <[e-mail]>
 *
 * This file is part of FFmpeg.
 *
 * FFmpeg is free software; you can redistribute it and/or
 * modify it under the terms of the GNU Lesser General Public
 * License as published by the Free Software Foundation; either
 * version 2.1 of the License, or (at your option) any later version.
 *
 * FFmpeg is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * Lesser General Public License for more details.
 *
 * You should have received a copy of the GNU Lesser General Public
 * License along with FFmpeg; if not, write to the Free Software
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 */

/**
 * @file libavutil/common.h
 * common internal and external API header
 */

#ifndef AVUTIL_COMMON_H
#define AVUTIL_COMMON_H

#include <ctype.h>
#include <errno.h>
#include <inttypes.h>
#include <limits.h>
#include <math.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>

#ifdef __GNUC__
#    define AV_GCC_VERSION_AT_LEAST(x,y) (__GNUC__ > x || __GNUC__ == x && __GNUC_MINOR__ >= y)
#else
#    define AV_GCC_VERSION_AT_LEAST(x,y) 0
#endif

#ifndef av_always_inline
#if AV_GCC_VERSION_AT_LEAST(3,1)
#    define av_always_inline __attribute__((always_inline)) inline
#else
#    define av_always_inline inline
#endif
#endif

#ifndef av_noinline
#if AV_GCC_VERSION_AT_LEAST(3,1)
#    define av_noinline __attribute__((noinline))
#else
#    define av_noinline
#endif
#endif

#ifndef av_pure
#if AV_GCC_VERSION_AT_LEAST(3,1)
#    define av_pure __attribute__((pure))
#else
#    define av_pure
#endif
#endif

#ifndef av_const
#if AV_GCC_VERSION_AT_LEAST(2,6)
#    define av_const __attribute__((const))
#else
#    define av_const
#endif
#endif

#ifndef av_cold
#if (!defined(__ICC) || __ICC > 1100) && AV_GCC_VERSION_AT_LEAST(4,3)
#    define av_cold __attribute__((cold))
#else
#    define av_cold
#endif
#endif

#ifndef av_flatten
#if AV_GCC_VERSION_AT_LEAST(4,1)
#    define av_flatten __attribute__((flatten))
#else
#    define av_flatten
#endif
#endif

#ifndef attribute_deprecated
#if AV_GCC_VERSION_AT_LEAST(3,1)
#    define attribute_deprecated __attribute__((deprecated))
#else
#    define attribute_deprecated
#endif
#endif

#ifndef av_unused
#if defined(__GNUC__)
#    define av_unused __attribute__((unused))
#else
#    define av_unused
#endif
#endif

#ifndef av_uninit
#if defined(__GNUC__) && !defined(__ICC)
#    define av_uninit(x) x=x
#else
#    define av_uninit(x) x
#endif
#endif

//rounded division & shift
#define RSHIFT(a,b) ((a) > 0 ? ((a) + ((1<<(b))>>1))>>(b) : ((a) + ((1<<(b))>>1)-1)>>(b))
/* assume b>0 */
#define ROUNDED_DIV(a,b) (((a)>0 ? (a) + ((b)>>1) : (a) - ((b)>>1))/(b))
#define FFABS(a) ((a) >= 0 ? (a) : (-(a)))
#define FFSIGN(a) ((a) > 0 ? 1 : -1)

#define FFMAX(a,b) ((a) > (b) ? (a) : (b))
#define FFMAX3(a,b,c) FFMAX(FFMAX(a,b),c)
#define FFMIN(a,b) ((a) > (b) ? (b) : (a))
#define FFMIN3(a,b,c) FFMIN(FFMIN(a,b),c)

#define FFSWAP(type,a,b) do{type SWAP_tmp= b; b= a; a= SWAP_tmp;}while(0)
#define FF_ARRAY_ELEMS(a) (sizeof(a) / sizeof((a)[0]))

/* misc math functions */
extern const uint8_t ff_log2_tab[256];

static inline av_const int av_log2(unsigned int v)
{
    int n = 0;
    if (v & 0xffff0000) {
        v >>= 16;
        n += 16;
    }
    if (v & 0xff00) {
        v >>= 8;
        n += 8;
    }
    n += ff_log2_tab[v];

    return n;
}

static inline av_const int av_log2_16bit(unsigned int v)
{
    int n = 0;
    if (v & 0xff00) {
        v >>= 8;
        n += 8;
    }
    n += ff_log2_tab[v];

    return n;
}

/**
 * Clips a signed integer value into the amin-amax range.
 * @param a value to clip
 * @param amin minimum value of the clip range
 * @param amax maximum value of the clip range
 * @return clipped value
 */
static inline av_const int av_clip(int a, int amin, int amax)
{
    if      (a < amin) return amin;
    else if (a > amax) return amax;
    else               return a;
}

/**
 * Clips a signed integer value into the 0-255 range.
 * @param a value to clip
 * @return clipped value
 */
static inline av_const uint8_t av_clip_uint8(int a)
{
    if (a&(~255)) return (-a)>>31;
    else          return a;
}

/**
 * Clips a signed integer value into the -32768,32767 range.
 * @param a value to clip
 * @return clipped value
 */
static inline av_const int16_t av_clip_int16(int a)
{
    if ((a+32768) & ~65535) return (a>>31) ^ 32767;
    else                    return a;
}

/**
 * Clips a float value into the amin-amax range.
 * @param a value to clip
 * @param amin minimum value of the clip range
 * @param amax maximum value of the clip range
 * @return clipped value
 */
static inline av_const float av_clipf(float a, float amin, float amax)
{
    if      (a < amin) return amin;
    else if (a > amax) return amax;
    else               return a;
}

#define MKTAG(a,b,c,d) (a | (b << 8) | (c << 16) | (d << 24))
#define MKBETAG(a,b,c,d) (d | (c << 8) | (b << 16) | (a << 24))

/*!
 * \def GET_UTF8(val, GET_BYTE, ERROR)
 * Converts a UTF-8 character (up to 4 bytes long) to its 32-bit UCS-4 encoded form
 * \param val is the output and should be of type uint32_t. It holds the converted
 * UCS-4 character and should be a left value.
 * \param GET_BYTE gets UTF-8 encoded bytes from any proper source. It can be
 * a function or a statement whose return value or evaluated value is of type
 * uint8_t. It will be executed up to 4 times for values in the valid UTF-8 range,
 * and up to 7 times in the general case.
 * \param ERROR action that should be taken when an invalid UTF-8 byte is returned
 * from GET_BYTE. It should be a statement that jumps out of the macro,
 * like exit(), goto, return, break, or continue.
 */
#define GET_UTF8(val, GET_BYTE, ERROR)\
    val= GET_BYTE;\
    {\
        int ones= 7 - av_log2(val ^ 255);\
        if(ones==1)\
            ERROR\
        val&= 127>>ones;\
        while(--ones > 0){\
            int tmp= GET_BYTE - 128;\
            if(tmp>>6)\
                ERROR\
            val= (val<<6) + tmp;\
        }\
    }

/*!
 * \def PUT_UTF8(val, tmp, PUT_BYTE)
 * Converts a 32-bit Unicode character to its UTF-8 encoded form (up to 4 bytes long).
 * \param val is an input-only argument and should be of type uint32_t. It holds
 * a UCS-4 encoded Unicode character that is to be converted to UTF-8. If
 * val is given as a function it is executed only once.
 * \param tmp is a temporary variable and should be of type uint8_t. It
 * represents an intermediate value during conversion that is to be
 * output by PUT_BYTE.
 * \param PUT_BYTE writes the converted UTF-8 bytes to any proper destination.
 * It could be a function or a statement, and uses tmp as the input byte.
 * For example, PUT_BYTE could be "*output++ = tmp;" PUT_BYTE will be
 * executed up to 4 times for values in the valid UTF-8 range and up to
 * 7 times in the general case, depending on the length of the converted
 * Unicode character.
 */
#define PUT_UTF8(val, tmp, PUT_BYTE)\
    {\
        int bytes, shift;\
        uint32_t in = val;\
        if (in < 0x80) {\
            tmp = in;\
            PUT_BYTE\
        } else {\
            bytes = (av_log2(in) + 4) / 5;\
            shift = (bytes - 1) * 6;\
            tmp = (256 - (256 >> bytes)) | (in >> shift);\
            PUT_BYTE\
            while (shift >= 6) {\
                shift -= 6;\
                tmp = 0x80 | ((in >> shift) & 0x3f);\
                PUT_BYTE\
            }\
        }\
    }

#include "mem.h"

#ifdef HAVE_AV_CONFIG_H
#    include "config.h"
#    include "internal.h"
#endif /* HAVE_AV_CONFIG_H */

#endif /* AVUTIL_COMMON_H */