mirror of
https://github.com/cemu-project/Cemu.git
synced 2024-11-23 01:29:19 +01:00
399 lines
13 KiB
ArmAsm
399 lines
13 KiB
ArmAsm
|
//******************************************************************************
|
||
|
//*
|
||
|
//* Copyright (C) 2015 The Android Open Source Project
|
||
|
//*
|
||
|
//* 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.
|
||
|
//*
|
||
|
//*****************************************************************************
|
||
|
//* Originally developed and contributed by Ittiam Systems Pvt. Ltd, Bangalore
|
||
|
//*/
|
||
|
///**
|
||
|
//******************************************************************************
|
||
|
//* @file
|
||
|
//* ih264_inter_pred_chroma_av8.s
|
||
|
//*
|
||
|
//* @brief
|
||
|
//* Contains function definitions for inter prediction interpolation.
|
||
|
//*
|
||
|
//* @author
|
||
|
//* Ittaim
|
||
|
//*
|
||
|
//* @par List of Functions:
|
||
|
//*
|
||
|
//* - ih264_inter_pred_chroma_av8()
|
||
|
//*
|
||
|
//* @remarks
|
||
|
//* None
|
||
|
//*
|
||
|
//*******************************************************************************
|
||
|
//*/
|
||
|
|
||
|
///* All the functions here are replicated from ih264_inter_pred_filters.c
|
||
|
//
|
||
|
|
||
|
///**
|
||
|
///**
|
||
|
///**
|
||
|
//
|
||
|
///**
|
||
|
//*******************************************************************************
|
||
|
//*
|
||
|
//* @brief
|
||
|
//* Interprediction chroma filter
|
||
|
//*
|
||
|
//* @par Description:
|
||
|
//* Applies filtering to chroma samples as mentioned in
|
||
|
//* sec 8.4.2.2.2 titled "chroma sample interpolation process"
|
||
|
//*
|
||
|
//* @param[in] pu1_src
|
||
|
//* UWORD8 pointer to the source containing alternate U and V samples
|
||
|
//*
|
||
|
//* @param[out] pu1_dst
|
||
|
//* UWORD8 pointer to the destination
|
||
|
//*
|
||
|
//* @param[in] src_strd
|
||
|
//* integer source stride
|
||
|
//*
|
||
|
//* @param[in] dst_strd
|
||
|
//* integer destination stride
|
||
|
//*
|
||
|
//* @param[in]uc_dx
|
||
|
//* dx value where the sample is to be produced(refer sec 8.4.2.2.2 )
|
||
|
//*
|
||
|
//* @param[in] uc_dy
|
||
|
//* dy value where the sample is to be produced(refer sec 8.4.2.2.2 )
|
||
|
//*
|
||
|
//* @param[in] ht
|
||
|
//* integer height of the array
|
||
|
//*
|
||
|
//* @param[in] wd
|
||
|
//* integer width of the array
|
||
|
//*
|
||
|
//* @returns
|
||
|
//*
|
||
|
//* @remarks
|
||
|
//* None
|
||
|
//*
|
||
|
//*******************************************************************************
|
||
|
//*/
|
||
|
|
||
|
//void ih264_inter_pred_chroma(UWORD8 *pu1_src,
|
||
|
// UWORD8 *pu1_dst,
|
||
|
// WORD32 src_strd,
|
||
|
// WORD32 dst_strd,
|
||
|
// WORD32 u1_dx,
|
||
|
// WORD32 u1_dy,
|
||
|
// WORD32 ht,
|
||
|
// WORD32 wd)
|
||
|
//**************Variables Vs Registers*****************************************
|
||
|
// x0 => *pu1_src
|
||
|
// x1 => *pu1_dst
|
||
|
// w2 => src_strd
|
||
|
// w3 => dst_strd
|
||
|
// w4 => u1_dx
|
||
|
// w5 => u1_dy
|
||
|
// w6 => height
|
||
|
// w7 => width
|
||
|
//
|
||
|
.text
|
||
|
.p2align 2
|
||
|
.include "ih264_neon_macros.s"
|
||
|
|
||
|
|
||
|
|
||
|
.global ih264_inter_pred_chroma_av8
|
||
|
|
||
|
ih264_inter_pred_chroma_av8:
|
||
|
|
||
|
|
||
|
|
||
|
// STMFD sp!, {x4-x12, x14} //store register values to stack
|
||
|
push_v_regs
|
||
|
stp x19, x20, [sp, #-16]!
|
||
|
sxtw x2, w2
|
||
|
sxtw x3, w3
|
||
|
sxtw x4, w4
|
||
|
sxtw x5, w5
|
||
|
sxtw x6, w6
|
||
|
sxtw x7, w7
|
||
|
|
||
|
|
||
|
|
||
|
|
||
|
|
||
|
sub x20, x4, #8 //8-u1_dx
|
||
|
neg x8, x20
|
||
|
sub x20, x5, #8 //8-u1_dy
|
||
|
neg x9, x20
|
||
|
mul x10, x8, x9 //
|
||
|
mul x11, x4, x9 //
|
||
|
|
||
|
dup v28.8b, w10
|
||
|
dup v29.8b, w11
|
||
|
|
||
|
mul x10, x8, x5 //
|
||
|
mul x11, x4, x5 //
|
||
|
|
||
|
dup v30.8b, w10
|
||
|
dup v31.8b, w11
|
||
|
|
||
|
subs x12, x7, #2 //if wd=4 branch to loop_4
|
||
|
beq loop_2
|
||
|
subs x12, x7, #4 //if wd=8 branch to loop_8
|
||
|
beq loop_4
|
||
|
|
||
|
loop_8:
|
||
|
ld1 {v0.8b, v1.8b, v2.8b}, [x0], x2 //// Load row0 ;
|
||
|
ext v3.8b, v0.8b , v1.8b , #2
|
||
|
ld1 {v5.8b, v6.8b, v7.8b}, [x0], x2 //// Load row1;
|
||
|
umull v20.8h, v0.8b, v28.8b
|
||
|
ext v8.8b, v5.8b , v6.8b , #2
|
||
|
umlal v20.8h, v3.8b, v29.8b
|
||
|
ext v9.8b, v6.8b , v7.8b , #2
|
||
|
umlal v20.8h, v5.8b, v30.8b
|
||
|
ext v4.8b, v1.8b , v2.8b , #2
|
||
|
umlal v20.8h, v8.8b, v31.8b
|
||
|
sqrshrun v26.8b, v20.8h, #6
|
||
|
umull v22.8h, v1.8b, v28.8b
|
||
|
ld1 {v10.8b, v11.8b, v12.8b}, [x0], x2 //// Load row2 ;
|
||
|
umlal v22.8h, v4.8b, v29.8b
|
||
|
ext v13.8b, v10.8b , v11.8b , #2
|
||
|
umlal v22.8h, v6.8b, v30.8b
|
||
|
ext v14.8b, v11.8b , v12.8b , #2
|
||
|
umlal v22.8h, v9.8b, v31.8b
|
||
|
sqrshrun v27.8b, v22.8h, #6
|
||
|
umull v24.8h, v5.8b, v28.8b
|
||
|
st1 { v26.8b, v27.8b}, [x1], x3 ////Store dest row
|
||
|
umlal v24.8h, v8.8b, v29.8b
|
||
|
ld1 {v0.8b, v1.8b, v2.8b}, [x0], x2 //// Load row3 ;
|
||
|
umlal v24.8h, v10.8b, v30.8b
|
||
|
ext v3.8b, v0.8b , v1.8b , #2
|
||
|
umlal v24.8h, v13.8b, v31.8b
|
||
|
ext v4.8b, v1.8b , v2.8b , #2
|
||
|
umull v16.8h, v6.8b, v28.8b
|
||
|
sqrshrun v18.8b, v24.8h, #6
|
||
|
umlal v16.8h, v9.8b, v29.8b
|
||
|
umlal v16.8h, v11.8b, v30.8b
|
||
|
umlal v16.8h, v14.8b, v31.8b
|
||
|
sqrshrun v19.8b, v16.8h, #6
|
||
|
st1 {v18.8b, v19.8b}, [x1], x3 // store row 1
|
||
|
umull v20.8h, v10.8b, v28.8b
|
||
|
umlal v20.8h, v13.8b, v29.8b
|
||
|
umlal v20.8h, v0.8b, v30.8b
|
||
|
umlal v20.8h, v3.8b, v31.8b
|
||
|
sqrshrun v26.8b, v20.8h, #6
|
||
|
umull v24.8h, v11.8b, v28.8b
|
||
|
ld1 {v5.8b, v6.8b, v7.8b}, [x0], x2 //// Load row4;
|
||
|
umlal v24.8h, v14.8b, v29.8b
|
||
|
ext v8.8b, v5.8b , v6.8b , #2
|
||
|
umlal v24.8h, v1.8b, v30.8b
|
||
|
ext v9.8b, v6.8b , v7.8b , #2
|
||
|
umlal v24.8h, v4.8b, v31.8b
|
||
|
umull v20.8h, v0.8b, v28.8b
|
||
|
sqrshrun v27.8b, v24.8h, #6
|
||
|
umlal v20.8h, v3.8b, v29.8b
|
||
|
st1 { v26.8b, v27.8b}, [x1], x3 ////Store dest row2
|
||
|
umlal v20.8h, v5.8b, v30.8b
|
||
|
umlal v20.8h, v8.8b, v31.8b
|
||
|
umull v22.8h, v1.8b, v28.8b
|
||
|
umlal v22.8h, v4.8b, v29.8b
|
||
|
umlal v22.8h, v6.8b, v30.8b
|
||
|
sqrshrun v26.8b, v20.8h, #6
|
||
|
umlal v22.8h, v9.8b, v31.8b
|
||
|
subs x12, x6, #4
|
||
|
sqrshrun v27.8b, v22.8h, #6
|
||
|
st1 { v26.8b, v27.8b}, [x1], x3 ////Store dest row3
|
||
|
|
||
|
beq end_func //If ht=4
|
||
|
|
||
|
ld1 {v10.8b, v11.8b, v12.8b}, [x0], x2 //// Load row5
|
||
|
ext v13.8b, v10.8b , v11.8b , #2
|
||
|
umull v24.8h, v5.8b, v28.8b
|
||
|
ext v14.8b, v11.8b , v12.8b , #2
|
||
|
ld1 {v0.8b, v1.8b, v2.8b}, [x0], x2 //// Load row6;
|
||
|
umlal v24.8h, v8.8b, v29.8b
|
||
|
umlal v24.8h, v10.8b, v30.8b
|
||
|
umlal v24.8h, v13.8b, v31.8b
|
||
|
ext v3.8b, v0.8b , v1.8b , #2
|
||
|
umull v16.8h, v6.8b, v28.8b
|
||
|
sqrshrun v18.8b, v24.8h, #6
|
||
|
umlal v16.8h, v9.8b, v29.8b
|
||
|
umlal v16.8h, v11.8b, v30.8b
|
||
|
umlal v16.8h, v14.8b, v31.8b
|
||
|
ext v4.8b, v1.8b , v2.8b , #2
|
||
|
sqrshrun v19.8b, v16.8h, #6
|
||
|
st1 { v18.8b, v19.8b}, [x1], x3 // store row 4
|
||
|
umull v20.8h, v10.8b, v28.8b
|
||
|
umlal v20.8h, v13.8b, v29.8b
|
||
|
umlal v20.8h, v0.8b, v30.8b
|
||
|
umlal v20.8h, v3.8b, v31.8b
|
||
|
ld1 {v5.8b, v6.8b, v7.8b}, [x0], x2 //// Load row7;
|
||
|
sqrshrun v26.8b, v20.8h, #6
|
||
|
umull v24.8h, v11.8b, v28.8b
|
||
|
umlal v24.8h, v14.8b, v29.8b
|
||
|
ext v8.8b, v5.8b , v6.8b , #2
|
||
|
umlal v24.8h, v1.8b, v30.8b
|
||
|
umlal v24.8h, v4.8b, v31.8b
|
||
|
ext v9.8b, v6.8b , v7.8b , #2
|
||
|
sqrshrun v27.8b, v24.8h, #6
|
||
|
st1 {v26.8b, v27.8b}, [x1], x3 ////Store dest row5
|
||
|
umull v20.8h, v0.8b, v28.8b
|
||
|
umlal v20.8h, v3.8b, v29.8b
|
||
|
umlal v20.8h, v5.8b, v30.8b
|
||
|
umlal v20.8h, v8.8b, v31.8b
|
||
|
ld1 {v10.8b, v11.8b, v12.8b}, [x0], x2 //// Load row8 ;
|
||
|
sqrshrun v26.8b, v20.8h, #6
|
||
|
umull v22.8h, v1.8b, v28.8b
|
||
|
umlal v22.8h, v4.8b, v29.8b
|
||
|
umlal v22.8h, v6.8b, v30.8b
|
||
|
ext v13.8b, v10.8b , v11.8b , #2
|
||
|
umlal v22.8h, v9.8b, v31.8b
|
||
|
ext v14.8b, v11.8b , v12.8b , #2
|
||
|
sqrshrun v27.8b, v22.8h, #6
|
||
|
st1 { v26.8b, v27.8b}, [x1], x3 ////Store dest row6
|
||
|
umull v24.8h, v5.8b, v28.8b
|
||
|
umlal v24.8h, v8.8b, v29.8b
|
||
|
umlal v24.8h, v10.8b, v30.8b
|
||
|
umlal v24.8h, v13.8b, v31.8b
|
||
|
umull v16.8h, v6.8b, v28.8b
|
||
|
sqrshrun v18.8b, v24.8h, #6
|
||
|
umlal v16.8h, v9.8b, v29.8b
|
||
|
umlal v16.8h, v11.8b, v30.8b
|
||
|
umlal v16.8h, v14.8b, v31.8b
|
||
|
sqrshrun v19.8b, v16.8h, #6
|
||
|
st1 { v18.8b, v19.8b}, [x1], x3 // store row 7
|
||
|
b end_func
|
||
|
|
||
|
loop_4:
|
||
|
ld1 {v0.8b, v1.8b}, [x0], x2 //// Load row0 ;
|
||
|
ext v2.8b, v0.8b , v1.8b , #2
|
||
|
ld1 {v3.8b, v4.8b}, [x0], x2 //// Load row1;
|
||
|
ext v5.8b, v3.8b , v4.8b , #2
|
||
|
umull v20.8h, v0.8b, v28.8b
|
||
|
umlal v20.8h, v2.8b, v29.8b
|
||
|
umlal v20.8h, v3.8b, v30.8b
|
||
|
umlal v20.8h, v5.8b, v31.8b
|
||
|
ld1 {v6.8b, v7.8b}, [x0], x2 //// Load row2
|
||
|
sqrshrun v26.8b, v20.8h, #6
|
||
|
ext v8.8b, v6.8b , v7.8b , #2
|
||
|
st1 {v26.8b}, [x1], x3 ////Store dest row0
|
||
|
umull v22.8h, v3.8b, v28.8b
|
||
|
umlal v22.8h, v5.8b, v29.8b
|
||
|
umlal v22.8h, v6.8b, v30.8b
|
||
|
umlal v22.8h, v8.8b, v31.8b
|
||
|
subs x12, x6, #2
|
||
|
sqrshrun v27.8b, v22.8h, #6
|
||
|
st1 {v27.8b}, [x1], x3 ////Store dest row1
|
||
|
beq end_func //If ht=2
|
||
|
|
||
|
ld1 {v9.8b, v10.8b}, [x0], x2 //// Load row3;
|
||
|
ext v11.8b, v9.8b , v10.8b , #2
|
||
|
umull v24.8h, v6.8b, v28.8b
|
||
|
umlal v24.8h, v8.8b, v29.8b
|
||
|
umlal v24.8h, v9.8b, v30.8b
|
||
|
umlal v24.8h, v11.8b, v31.8b
|
||
|
ld1 {v0.8b, v1.8b}, [x0], x2 //// Load row4 ;
|
||
|
sqrshrun v16.8b, v24.8h, #6
|
||
|
ext v2.8b, v0.8b , v1.8b , #2
|
||
|
st1 {v16.8b}, [x1], x3 ////Store dest row2
|
||
|
umull v18.8h, v9.8b, v28.8b
|
||
|
umlal v18.8h, v11.8b, v29.8b
|
||
|
umlal v18.8h, v0.8b, v30.8b
|
||
|
umlal v18.8h, v2.8b, v31.8b
|
||
|
subs x12, x6, #4
|
||
|
sqrshrun v17.8b, v18.8h, #6
|
||
|
st1 {v17.8b}, [x1], x3 ////Store dest row3
|
||
|
beq end_func //If ht=4
|
||
|
|
||
|
ld1 {v3.8b, v4.8b}, [x0], x2 //// Load row5;
|
||
|
ext v5.8b, v3.8b , v4.8b , #2
|
||
|
umull v20.8h, v0.8b, v28.8b
|
||
|
umlal v20.8h, v2.8b, v29.8b
|
||
|
umlal v20.8h, v3.8b, v30.8b
|
||
|
umlal v20.8h, v5.8b, v31.8b
|
||
|
ld1 {v6.8b, v7.8b}, [x0], x2 //// Load row6 ;
|
||
|
sqrshrun v26.8b, v20.8h, #6
|
||
|
ext v8.8b, v6.8b , v7.8b , #2
|
||
|
st1 {v26.8b}, [x1], x3 ////Store dest row4
|
||
|
umull v22.8h, v3.8b, v28.8b
|
||
|
umlal v22.8h, v5.8b, v29.8b
|
||
|
umlal v22.8h, v6.8b, v30.8b
|
||
|
umlal v22.8h, v8.8b, v31.8b
|
||
|
ld1 {v9.8b, v10.8b}, [x0], x2 //// Load row7;
|
||
|
sqrshrun v27.8b, v22.8h, #6
|
||
|
ext v11.8b, v9.8b , v10.8b , #2
|
||
|
st1 {v27.8b}, [x1], x3 ////Store dest row5
|
||
|
umull v24.8h, v6.8b, v28.8b
|
||
|
umlal v24.8h, v8.8b, v29.8b
|
||
|
umlal v24.8h, v9.8b, v30.8b
|
||
|
umlal v24.8h, v11.8b, v31.8b
|
||
|
ld1 {v0.8b, v1.8b}, [x0], x2 //// Load row8;
|
||
|
sqrshrun v16.8b, v24.8h, #6
|
||
|
ext v2.8b, v0.8b , v1.8b , #2
|
||
|
st1 {v16.8b}, [x1], x3 ////Store dest row6
|
||
|
umull v18.8h, v9.8b, v28.8b
|
||
|
umlal v18.8h, v11.8b, v29.8b
|
||
|
umlal v18.8h, v0.8b, v30.8b
|
||
|
umlal v18.8h, v2.8b, v31.8b
|
||
|
sqrshrun v17.8b, v18.8h, #6
|
||
|
st1 {v17.8b}, [x1], x3 ////Store dest row7
|
||
|
b end_func
|
||
|
|
||
|
loop_2:
|
||
|
ld1 {v0.8b}, [x0], x2 //// Load row0 ;
|
||
|
ext v2.8b, v0.8b , v0.8b , #2
|
||
|
ld1 {v3.8b}, [x0], x2 //// Load row1;
|
||
|
ext v5.8b, v3.8b , v3.8b , #2
|
||
|
umull v20.8h, v0.8b, v28.8b
|
||
|
umlal v20.8h, v2.8b, v29.8b
|
||
|
umlal v20.8h, v3.8b, v30.8b
|
||
|
umlal v20.8h, v5.8b, v31.8b
|
||
|
ld1 {v6.8b}, [x0], x2 //// Load row2
|
||
|
sqrshrun v26.8b, v20.8h, #6
|
||
|
ext v8.8b, v6.8b , v6.8b , #2
|
||
|
st1 {v26.s}[0], [x1], x3 ////Store dest row0
|
||
|
umull v22.8h, v3.8b, v28.8b
|
||
|
umlal v22.8h, v5.8b, v29.8b
|
||
|
umlal v22.8h, v6.8b, v30.8b
|
||
|
umlal v22.8h, v8.8b, v31.8b
|
||
|
subs x12, x6, #2
|
||
|
sqrshrun v27.8b, v22.8h, #6
|
||
|
st1 {v27.s}[0], [x1], x3 ////Store dest row1
|
||
|
beq end_func //If ht=2
|
||
|
|
||
|
ld1 {v9.8b}, [x0], x2 //// Load row3;
|
||
|
ext v11.8b, v9.8b , v9.8b , #2
|
||
|
umull v24.8h, v6.8b, v28.8b
|
||
|
umlal v24.8h, v8.8b, v29.8b
|
||
|
umlal v24.8h, v9.8b, v30.8b
|
||
|
umlal v24.8h, v11.8b, v31.8b
|
||
|
ld1 {v0.8b}, [x0], x2 //// Load row4 ;
|
||
|
sqrshrun v16.8b, v24.8h, #6
|
||
|
ext v2.8b, v0.8b , v0.8b , #2
|
||
|
st1 {v16.s}[0], [x1], x3 ////Store dest row2
|
||
|
umull v18.8h, v9.8b, v28.8b
|
||
|
umlal v18.8h, v11.8b, v29.8b
|
||
|
umlal v18.8h, v0.8b, v30.8b
|
||
|
umlal v18.8h, v2.8b, v31.8b
|
||
|
sqrshrun v17.8b, v18.8h, #6
|
||
|
st1 {v17.s}[0], [x1], x3 ////Store dest row3
|
||
|
|
||
|
|
||
|
end_func:
|
||
|
// LDMFD sp!,{x4-x12,PC} //Restoring registers from stack
|
||
|
ldp x19, x20, [sp], #16
|
||
|
pop_v_regs
|
||
|
ret
|
||
|
|
||
|
|