yading@10: /*
yading@10:  * Copyright (c) 2010 Mans Rullgard
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: #include "libavutil/arm/asm.S"
yading@10: #include "asm-offsets.h"
yading@10: 
yading@10: function ff_dct_unquantize_h263_inter_neon, export=1
yading@10:         add             r12, r0,  #BLOCK_LAST_INDEX
yading@10:         ldr             r12, [r12, r2, lsl #2]
yading@10:         add             r0,  r0,  #INTER_SCANTAB_RASTER_END
yading@10:         ldrb            r12, [r0, r12]
yading@10:         sub             r2,  r3,  #1
yading@10:         lsl             r0,  r3,  #1
yading@10:         orr             r2,  r2,  #1
yading@10:         add             r3,  r12, #1
yading@10: endfunc
yading@10: 
yading@10: function ff_dct_unquantize_h263_neon, export=1
yading@10:         vdup.16         q15, r0                 @ qmul
yading@10:         vdup.16         q14, r2                 @ qadd
yading@10:         vneg.s16        q13, q14
yading@10:         cmp             r3,  #4
yading@10:         mov             r0,  r1
yading@10:         ble             2f
yading@10: 1:
yading@10:         vld1.16         {q0},     [r0,:128]!
yading@10:         vclt.s16        q3,  q0,  #0
yading@10:         vld1.16         {q8},     [r0,:128]!
yading@10:         vceq.s16        q1,  q0,  #0
yading@10:         vmul.s16        q2,  q0,  q15
yading@10:         vclt.s16        q11, q8,  #0
yading@10:         vmul.s16        q10, q8,  q15
yading@10:         vbsl            q3,  q13, q14
yading@10:         vbsl            q11, q13, q14
yading@10:         vadd.s16        q2,  q2,  q3
yading@10:         vceq.s16        q9,  q8,  #0
yading@10:         vadd.s16        q10, q10, q11
yading@10:         vbif            q0,  q2,  q1
yading@10:         vbif            q8,  q10, q9
yading@10:         subs            r3,  r3,  #16
yading@10:         vst1.16         {q0},     [r1,:128]!
yading@10:         vst1.16         {q8},     [r1,:128]!
yading@10:         it              le
yading@10:         bxle            lr
yading@10:         cmp             r3,  #8
yading@10:         bgt             1b
yading@10: 2:
yading@10:         vld1.16         {d0},     [r0,:64]
yading@10:         vclt.s16        d3,  d0,  #0
yading@10:         vceq.s16        d1,  d0,  #0
yading@10:         vmul.s16        d2,  d0,  d30
yading@10:         vbsl            d3,  d26, d28
yading@10:         vadd.s16        d2,  d2,  d3
yading@10:         vbif            d0,  d2,  d1
yading@10:         vst1.16         {d0},     [r1,:64]
yading@10:         bx              lr
yading@10: endfunc
yading@10: 
yading@10: function ff_dct_unquantize_h263_intra_neon, export=1
yading@10:         push            {r4-r6,lr}
yading@10:         add             r12, r0,  #BLOCK_LAST_INDEX
yading@10:         ldr             r6,  [r0, #AC_PRED]
yading@10:         add             lr,  r0,  #INTER_SCANTAB_RASTER_END
yading@10:         cmp             r6,  #0
yading@10:         it              ne
yading@10:         movne           r12, #63
yading@10:         bne             1f
yading@10:         ldr             r12, [r12, r2, lsl #2]
yading@10:         ldrb            r12, [lr, r12]
yading@10: 1:      ldr             r5,  [r0, #H263_AIC]
yading@10:         ldrsh           r4,  [r1]
yading@10:         cmp             r5,  #0
yading@10:         mov             r5,  r1
yading@10:         it              ne
yading@10:         movne           r2,  #0
yading@10:         bne             2f
yading@10:         cmp             r2,  #4
yading@10:         it              ge
yading@10:         addge           r0,  r0,  #4
yading@10:         sub             r2,  r3,  #1
yading@10:         ldr             r6,  [r0, #Y_DC_SCALE]
yading@10:         orr             r2,  r2,  #1
yading@10:         smulbb          r4,  r4,  r6
yading@10: 2:      lsl             r0,  r3,  #1
yading@10:         add             r3,  r12, #1
yading@10:         bl              ff_dct_unquantize_h263_neon
yading@10:         vmov.16         d0[0], r4
yading@10:         vst1.16         {d0[0]},  [r5]
yading@10:         pop             {r4-r6,pc}
yading@10: endfunc