yading@10: /*
yading@10:  * Copyright (C) 2010 Mans Rullgard <mans@mansr.com>
yading@10:  *
yading@10:  * This file is part of FFmpeg.
yading@10:  *
yading@10:  * FFmpeg is free software; you can redistribute it and/or
yading@10:  * modify it under the terms of the GNU Lesser General Public
yading@10:  * License as published by the Free Software Foundation; either
yading@10:  * version 2.1 of the License, or (at your option) any later version.
yading@10:  *
yading@10:  * FFmpeg is distributed in the hope that it will be useful,
yading@10:  * but WITHOUT ANY WARRANTY; without even the implied warranty of
yading@10:  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
yading@10:  * Lesser General Public License for more details.
yading@10:  *
yading@10:  * You should have received a copy of the GNU Lesser General Public
yading@10:  * License along with FFmpeg; if not, write to the Free Software
yading@10:  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
yading@10:  */
yading@10: 
yading@10: #ifndef AVCODEC_ARM_VP56_ARITH_H
yading@10: #define AVCODEC_ARM_VP56_ARITH_H
yading@10: 
yading@10: #if CONFIG_THUMB
yading@10: #   define A(x)
yading@10: #   define T(x) x
yading@10: #else
yading@10: #   define A(x) x
yading@10: #   define T(x)
yading@10: #endif
yading@10: 
yading@10: #if CONFIG_THUMB || defined __clang__
yading@10: #   define L(x)
yading@10: #   define U(x) x
yading@10: #else
yading@10: #   define L(x) x
yading@10: #   define U(x)
yading@10: #endif
yading@10: 
yading@10: #if HAVE_ARMV6_INLINE
yading@10: 
yading@10: #define vp56_rac_get_prob vp56_rac_get_prob_armv6
yading@10: static inline int vp56_rac_get_prob_armv6(VP56RangeCoder *c, int pr)
yading@10: {
yading@10:     unsigned shift     = ff_vp56_norm_shift[c->high];
yading@10:     unsigned code_word = c->code_word << shift;
yading@10:     unsigned high      = c->high << shift;
yading@10:     unsigned bit;
yading@10: 
yading@10:     __asm__ ("adds    %3,  %3,  %0           \n"
yading@10:              "itt     cs                     \n"
yading@10:              "cmpcs   %7,  %4                \n"
yading@10:            L("ldrcsh  %2,  [%4], #2          \n")
yading@10:            U("ldrhcs  %2,  [%4], #2          \n")
yading@10:              "rsb     %0,  %6,  #256         \n"
yading@10:              "smlabb  %0,  %5,  %6,  %0      \n"
yading@10:            T("itttt   cs                     \n")
yading@10:              "rev16cs %2,  %2                \n"
yading@10:            T("lslcs   %2,  %2,  %3           \n")
yading@10:            T("orrcs   %1,  %1,  %2           \n")
yading@10:            A("orrcs   %1,  %1,  %2,  lsl %3  \n")
yading@10:              "subcs   %3,  %3,  #16          \n"
yading@10:              "lsr     %0,  %0,  #8           \n"
yading@10:              "cmp     %1,  %0,  lsl #16      \n"
yading@10:              "ittte   ge                     \n"
yading@10:              "subge   %1,  %1,  %0,  lsl #16 \n"
yading@10:              "subge   %0,  %5,  %0           \n"
yading@10:              "movge   %2,  #1                \n"
yading@10:              "movlt   %2,  #0                \n"
yading@10:              : "=&r"(c->high), "=&r"(c->code_word), "=&r"(bit),
yading@10:                "+&r"(c->bits), "+&r"(c->buffer)
yading@10:              : "r"(high), "r"(pr), "r"(c->end - 1),
yading@10:                "0"(shift), "1"(code_word)
yading@10:              : "cc");
yading@10: 
yading@10:     return bit;
yading@10: }
yading@10: 
yading@10: #define vp56_rac_get_prob_branchy vp56_rac_get_prob_branchy_armv6
yading@10: static inline int vp56_rac_get_prob_branchy_armv6(VP56RangeCoder *c, int pr)
yading@10: {
yading@10:     unsigned shift     = ff_vp56_norm_shift[c->high];
yading@10:     unsigned code_word = c->code_word << shift;
yading@10:     unsigned high      = c->high << shift;
yading@10:     unsigned low;
yading@10:     unsigned tmp;
yading@10: 
yading@10:     __asm__ ("adds    %3,  %3,  %0           \n"
yading@10:              "itt     cs                     \n"
yading@10:              "cmpcs   %7,  %4                \n"
yading@10:            L("ldrcsh  %2,  [%4], #2          \n")
yading@10:            U("ldrhcs  %2,  [%4], #2          \n")
yading@10:              "rsb     %0,  %6,  #256         \n"
yading@10:              "smlabb  %0,  %5,  %6,  %0      \n"
yading@10:            T("itttt   cs                     \n")
yading@10:              "rev16cs %2,  %2                \n"
yading@10:            T("lslcs   %2,  %2,  %3           \n")
yading@10:            T("orrcs   %1,  %1,  %2           \n")
yading@10:            A("orrcs   %1,  %1,  %2,  lsl %3  \n")
yading@10:              "subcs   %3,  %3,  #16          \n"
yading@10:              "lsr     %0,  %0,  #8           \n"
yading@10:              "lsl     %2,  %0,  #16          \n"
yading@10:              : "=&r"(low), "+&r"(code_word), "=&r"(tmp),
yading@10:                "+&r"(c->bits), "+&r"(c->buffer)
yading@10:              : "r"(high), "r"(pr), "r"(c->end - 1), "0"(shift)
yading@10:              : "cc");
yading@10: 
yading@10:     if (code_word >= tmp) {
yading@10:         c->high      = high - low;
yading@10:         c->code_word = code_word - tmp;
yading@10:         return 1;
yading@10:     }
yading@10: 
yading@10:     c->high      = low;
yading@10:     c->code_word = code_word;
yading@10:     return 0;
yading@10: }
yading@10: 
yading@10: #endif
yading@10: 
yading@10: #endif /* AVCODEC_ARM_VP56_ARITH_H */