You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

238 lines
7.9 KiB

///*****************************************************************************
//*
//* Copyright (C) 2012 Ittiam Systems Pvt Ltd, Bangalore
//*
//* Licensed under the Apache License, Version 2.0 (the "License");
//* you may not use this file except in compliance with the License.
//* You may obtain a copy of the License at:
//*
//* http://www.apache.org/licenses/LICENSE-2.0
//*
//* Unless required by applicable law or agreed to in writing, software
//* distributed under the License is distributed on an "AS IS" BASIS,
//* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
//* See the License for the specific language governing permissions and
//* limitations under the License.
//*
//*****************************************************************************/
///**
// *******************************************************************************
// * @file
// * ihevc_itrans_recon_4x4_neon.s
// *
// * @brief
// * contains function definitions for single stage inverse transform
// *
// * @author
// * naveen sr
// *
// * @par list of functions:
// * - ihevc_itrans_recon_4x4()
// *
// * @remarks
// * none
// *
// *******************************************************************************
//*/
// /**
// *******************************************************************************
// *
// * @brief
// * this function performs inverse transform and reconstruction for 4x4
// * input block
// *
// * @par description:
// * performs inverse transform and adds the prediction data and clips output
// * to 8 bit
// *
// * @param[in] pi2_src
// * input 4x4 coefficients
// *
// * @param[in] pi2_tmp
// * temporary 4x4 buffer for storing inverse
// *
// * transform
// * 1st stage output
// *
// * @param[in] pu1_pred
// * prediction 4x4 block
// *
// * @param[out] pu1_dst
// * output 4x4 block
// *
// * @param[in] src_strd
// * input stride
// *
// * @param[in] pred_strd
// * prediction stride
// *
// * @param[in] dst_strd
// * output stride
// *
// * @param[in] shift
// * output shift
// *
// * @param[in] zero_cols
// * zero columns in pi2_src
// *
// * @returns void
// *
// * @remarks
// * none
// *
// *******************************************************************************
// */
//void ihevc_itrans_recon_4x4(word16 *pi2_src,
// word16 *pi2_tmp,
// uword8 *pu1_pred,
// uword8 *pu1_dst,
// word32 src_strd,
// word32 pred_strd,
// word32 dst_strd,
// word32 zero_cols)
//**************variables vs registers*************************
// x0 => *pi2_src
// x1 => *pi2_tmp
// x2 => *pu1_pred
// x3 => *pu1_dst
// x4 => src_strd
// x5 => pred_strd
// x6 => dst_strd
// x7 => zero_cols
.text
.align 4
.include "ihevc_neon_macros.s"
.set shift_stage1_idct , 7
.set shift_stage2_idct , 12
.globl ihevc_itrans_recon_4x4_av8
.extern g_ai2_ihevc_trans_4_transpose
.type ihevc_itrans_recon_4x4_av8, %function
ihevc_itrans_recon_4x4_av8:
// stmfd sp!, {x4-x12, x14} //stack stores the values of the arguments
stp x19, x20,[sp,#-16]!
adrp x8, :got:g_ai2_ihevc_trans_4_transpose
ldr x8, [x8, #:got_lo12:g_ai2_ihevc_trans_4_transpose]
add x4,x4,x4 // src_strd in terms of word16
add x9,x0,x4 // pi2_src[0] + src_strd
ld1 {v4.4h},[x8] //loading first row of g_ai2_ihevc_trans_4_transpose
// d4 = {36,64,83,64}
//index = 3 2 1 0
add x10,x9,x4, lsl #1 // 3*src_strd
add x4,x4,x4
ld1 {v1.4h},[x9] //loading pi2_src 2nd row
ld1 {v3.4h},[x10] //loading pi2_src 4th row
ld1 {v0.4h},[x0],x4 //loading pi2_src 1st row
ld1 {v2.4h},[x0],x4 //loading pi2_src 3rd row
// first stage computation starts
smull v6.4s, v1.4h, v4.h[1] //83 * pi2_src[1]
smlal v6.4s, v3.4h, v4.h[3] //o[0] = 83 * pi2_src[1] + 36 * pi2_src[3]
smull v5.4s, v1.4h, v4.h[3] //36 * pi2_src[1]
ld1 {v22.s}[0],[x2],x5
smlsl v5.4s, v3.4h, v4.h[1] //o[1] = 36 * pi2_src[1] - 83 * pi2_src[3]
saddl v7.4s, v0.4h, v2.4h //pi2_src[0] + pi2_src[2]
ssubl v17.4s, v0.4h, v2.4h //pi2_src[0] - pi2_src[2]
shl v7.4s, v7.4s,#6 //e[0] = 64*(pi2_src[0] + pi2_src[2])
shl v17.4s, v17.4s,#6 //e[1] = 64*(pi2_src[0] - pi2_src[2])
add v19.4s, v7.4s , v6.4s //((e[0] + o[0] )
add v16.4s, v17.4s , v5.4s //((e[1] + o[1])
sub v18.4s, v17.4s , v5.4s //((e[1] - o[1])
sub v20.4s, v7.4s , v6.4s //((e[0] - o[0])
sqrshrn v28.4h, v19.4s,#shift_stage1_idct //pi2_out[0] = clip_s16((e[0] + o[0] + add)>>shift) )
sqrshrn v29.4h, v16.4s,#shift_stage1_idct //pi2_out[1] = clip_s16((e[1] + o[1] + add)>>shift) )
sqrshrn v30.4h, v18.4s,#shift_stage1_idct //pi2_out[2] = clip_s16((e[0] - o[0] + add)>>shift) )
sqrshrn v31.4h, v20.4s,#shift_stage1_idct //pi2_out[3] = clip_s16((e[0] - o[0] + add)>>shift) )
trn1 v24.4h, v28.4h, v29.4h
trn2 v25.4h, v28.4h, v29.4h
trn1 v26.4h, v30.4h, v31.4h
trn2 v27.4h, v30.4h, v31.4h
trn1 v0.2s, v24.2s, v26.2s
trn2 v2.2s, v24.2s, v26.2s
trn1 v1.2s, v25.2s, v27.2s
trn2 v3.2s, v25.2s, v27.2s
// first stage ends
// output in d0,d1,d2,d3
// second stage starts
smull v6.4s, v1.4h, v4.h[1] //83 * pi2_src[1]
ld1 {v22.s}[1],[x2],x5
smlal v6.4s, v3.4h, v4.h[3] //o[0] = 83 * pi2_src[1] + 36 * pi2_src[3]
smull v5.4s, v1.4h, v4.h[3] //36 * pi2_src[1]
smlsl v5.4s, v3.4h, v4.h[1] //o[1] = 36 * pi2_src[1] - 83 * pi2_src[3]
ld1 {v23.s}[0],[x2],x5
saddl v7.4s, v0.4h, v2.4h //pi2_src[0] + pi2_src[2]
ssubl v17.4s, v0.4h, v2.4h //pi2_src[0] - pi2_src[2]
shl v7.4s, v7.4s,#6 //e[0] = 64*(pi2_src[0] + pi2_src[2])
shl v17.4s, v17.4s,#6 //e[1] = 64*(pi2_src[0] - pi2_src[2])
add v19.4s, v7.4s , v6.4s //((e[0] + o[0] )
add v16.4s, v17.4s , v5.4s //((e[1] + o[1])
sub v18.4s, v17.4s , v5.4s //((e[1] - o[1])
sub v20.4s, v7.4s , v6.4s //((e[0] - o[0])
sqrshrn v28.4h, v19.4s,#shift_stage2_idct //pi2_out[0] = clip_s16((e[0] + o[0] + add)>>shift) )
sqrshrn v29.4h, v16.4s,#shift_stage2_idct //pi2_out[1] = clip_s16((e[1] + o[1] + add)>>shift) )
sqrshrn v30.4h, v18.4s,#shift_stage2_idct //pi2_out[2] = clip_s16((e[0] - o[0] + add)>>shift) )
sqrshrn v31.4h, v20.4s,#shift_stage2_idct //pi2_out[3] = clip_s16((e[0] - o[0] + add)>>shift) )
ld1 {v23.s}[1],[x2],x5
trn1 v24.4h, v28.4h, v29.4h
trn2 v25.4h, v28.4h, v29.4h
trn1 v26.4h, v30.4h, v31.4h
trn2 v27.4h, v30.4h, v31.4h
trn1 v0.2s, v24.2s, v26.2s
trn2 v2.2s, v24.2s, v26.2s
trn1 v1.2s, v25.2s, v27.2s
trn2 v3.2s, v25.2s, v27.2s
// second stage ends
// output in d0,d1,d2,d3
// second stage computation ends
// loading pred
mov v0.d[1],v1.d[0]
mov v2.d[1],v3.d[0]
uaddw v0.8h, v0.8h , v22.8b // pi2_out(16bit) + pu1_pred(8bit)
uaddw v2.8h, v2.8h , v23.8b // pi2_out(16bit) + pu1_pred(8bit)
sqxtun v0.8b, v0.8h // clip_u8(pi2_out(16bit) + pu1_pred(8bit))
sqxtun v1.8b, v2.8h // clip_u8(pi2_out(16bit) + pu1_pred(8bit))
// storing destination
st1 {v0.s}[0],[x3],x6
st1 {v0.s}[1],[x3],x6
st1 {v1.s}[0],[x3],x6
st1 {v1.s}[1],[x3],x6
// ldmfd sp!,{x4-x12,x15} //reload the registers from sp
ldp x19, x20,[sp],#16
ret