wrapper/libhevc-mirror
1
0
Fork 0
mirror of https://github.com/ittiam-systems/libhevc.git synced 2026-04-03 04:40:49 +07:00
Code Issues Projects Releases Packages Wiki Activity Actions 1

Compare commits

base: wrapper:main
Branches Tags
wrapper:main
wrapper:v1.7.0
wrapper:v1.6.0
wrapper:v1.5.1
wrapper:v1.5.0
wrapper:v1.4.0
wrapper:v1.3.0
wrapper:v1.2.0
wrapper:v1.1.1
wrapper:v1.1.0
..
compare: wrapper:v1.5.0
Branches Tags
wrapper:main
wrapper:v1.7.0
wrapper:v1.6.0
wrapper:v1.5.1
wrapper:v1.5.0
wrapper:v1.4.0
wrapper:v1.3.0
wrapper:v1.2.0
wrapper:v1.1.1
wrapper:v1.1.0

No commits in common. "main" and "v1.5.0" have entirely different histories.
main ... v1.5.0

165 changed files with 3844 additions and 10464 deletions
Download patch file Download diff file Expand all files Collapse all files

2
.github/workflows/cifuzz.yml vendored
View file

@ -20,7 +20,7 @@ jobs:
language: c++
fuzz-seconds: 600
- name: Upload Crash
uses: actions/upload-artifact@v4
uses: actions/upload-artifact@v3
if: failure() && steps.build.outcome == 'success'
with:
name: artifacts

88
.github/workflows/cmake.yml vendored
View file

@ -2,100 +2,22 @@ name: CMake
on:
push:
branches: [ "main" ]
pull_request:
branches: [ "main" ]
env:
BUILD_TYPE: Release
jobs:
build:
strategy:
matrix:
include:
- name: ubuntu-latest-gcc-cmake
os: ubuntu-latest
cc: gcc
cxx: g++
build-system: cmake
cmake-opts: ''
- name: ubuntu-latest-clang-cmake
os: ubuntu-latest
cc: clang
cxx: clang++
build-system: cmake
cmake-opts: ''
- name: ubuntu-24.04-arm-clang-cmake
os: ubuntu-24.04-arm
cc: clang
cxx: clang++
build-system: cmake
cmake-opts: ''
- name: ubuntu-latest-clang-cmake-asan-fuzzer
os: ubuntu-latest
cc: clang
cxx: clang++
build-system: cmake
cmake-opts: '-DSANITIZE=fuzzer-no-link,address'
- name: ubuntu-latest-clang-cmake-ninja
os: ubuntu-latest
cc: clang
cxx: clang++
build-system: cmake
cmake-opts: '-G Ninja'
- name: macos-latest-clang-cmake
os: macos-latest
cc: clang
cxx: clang++
build-system: cmake
cmake-opts: ''
- name: ubuntu-latest-cross-aarch64-cmake
os: ubuntu-latest
cc: aarch64-linux-gnu-gcc
cxx: aarch64-linux-gnu-g++
build-system: cmake
cmake-opts: '-DCMAKE_TOOLCHAIN_FILE=../cmake/toolchains/aarch64_toolchain.cmake'
- name: ubuntu-latest-cross-aarch32-cmake
os: ubuntu-latest
cc: arm-linux-gnueabihf-gcc
cxx: arm-linux-gnueabihf-g++
build-system: cmake
cmake-opts: '-DCMAKE_TOOLCHAIN_FILE=../cmake/toolchains/aarch32_toolchain.cmake'
runs-on: ${{ matrix.os }}
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- name: Install Linux dependencies
if: startsWith(matrix.os,'ubuntu') && contains(matrix.cmake-opts,'-G Ninja')
run: |
sudo apt-get update
sudo apt-get install -y ninja-build
- name: Install cross-aarch64 dependencies
if: startsWith(matrix.os,'ubuntu') && contains(matrix.cmake-opts,'aarch64')
run: |
sudo apt-get update
sudo apt-get install gcc-aarch64-linux-gnu g++-aarch64-linux-gnu
- name: Install cross-arm dependencies
if: startsWith(matrix.os,'ubuntu') && contains(matrix.cmake-opts,'aarch32')
run: |
sudo apt-get update
sudo apt-get install gcc-arm-linux-gnueabihf g++-arm-linux-gnueabihf
- uses: actions/checkout@v3
- name: Configure CMake
env:
CC: ${{ matrix.cc }}
CXX: ${{ matrix.cxx }}
run: cmake -B ${{github.workspace}}/out -DCMAKE_BUILD_TYPE=${{env.BUILD_TYPE}} ${{ matrix.cmake-opts }}
run: cmake -B ${{github.workspace}}/out -DCMAKE_BUILD_TYPE=${{env.BUILD_TYPE}} -DCMAKE_C_COMPILER=clang -DCMAKE_CXX_COMPILER=clang++
- name: Build
run: cmake --build ${{github.workspace}}/out --config ${{env.BUILD_TYPE}}

18
.vscode/c_cpp_properties.json vendored
View file

@ -1,18 +0,0 @@
{
"configurations": [
{
"name": "Native",
"includePath": [
"${workspaceFolder}/**",
"${workspaceFolder}/common",
"${workspaceFolder}/decoder",
"${workspaceFolder}/encoder"
],
"defines": [],
"cStandard": "c17",
"cppStandard": "c++17",
"configurationProvider": "ms-vscode.cmake-tools"
}
],
"version": 4
}

81
.vscode/launch.json vendored
View file

@ -1,81 +0,0 @@
{
"version": "0.2.0",
"configurations": [
{
"name": "Run hevcenc - Linux",
"type": "cppdbg",
"request": "launch",
"program": "${workspaceFolder}/build/hevcenc",
"args": ["../test/encoder/vid_enc_cfg.txt"],
"stopAtEntry": false,
"cwd": "${workspaceFolder}/build",
"environment": [],
"externalConsole": false,
"MIMode": "gdb",
"setupCommands": [
{
"description": "Enable pretty-printing for gdb",
"text": "-enable-pretty-printing",
"ignoreFailures": true
}
]
},
{
"name": "Run hevcdec - Linux",
"type": "cppdbg",
"request": "launch",
"program": "${workspaceFolder}/build/hevcdec",
"args": ["../test/decoder/test.cfg"],
"stopAtEntry": false,
"cwd": "${workspaceFolder}/build",
"environment": [],
"externalConsole": false,
"MIMode": "gdb",
"setupCommands": [
{
"description": "Enable pretty-printing for gdb",
"text": "-enable-pretty-printing",
"ignoreFailures": true
}
]
},
{
"name": "Run hevcenc - Mac",
"type": "cppdbg",
"request": "launch",
"program": "${workspaceFolder}/build/hevcenc",
"args": ["../test/encoder/vid_enc_cfg.txt"],
"stopAtEntry": false,
"cwd": "${workspaceFolder}/build",
"environment": [],
"externalConsole": false,
"MIMode": "lldb",
"setupCommands": [
{
"description": "Enable pretty-printing for lldb",
"text": "-enable-pretty-printing",
"ignoreFailures": true
}
]
},
{
"name": "Run hevcdec - Mac",
"type": "cppdbg",
"request": "launch",
"program": "${workspaceFolder}/build/hevcdec",
"args": ["../test/decoder/test.cfg"],
"stopAtEntry": false,
"cwd": "${workspaceFolder}/build",
"environment": [],
"externalConsole": false,
"MIMode": "lldb",
"setupCommands": [
{
"description": "Enable pretty-printing for lldb",
"text": "-enable-pretty-printing",
"ignoreFailures": true
}
]
}
]
}

19
.vscode/settings.json vendored
View file

@ -1,19 +0,0 @@
{
"cmake.buildDirectory": "${workspaceFolder}/build",
"cmake.sourceDirectory": "${workspaceFolder}",
"cmake.configureArgs": [
"-DENABLE_MVC=OFF",
"-DENABLE_SVC=OFF",
"-DENABLE_TESTS=OFF",
"-DCMAKE_C_COMPILER=clang",
"-DCMAKE_CXX_COMPILER=clang++"
],
"cmake.preferredGenerators": [
"Unix Makefiles"
],
"cmake.debugConfig": {
"hevcenc": "hevcenc",
"hevcdec": "hevcdec"
},
"C_Cpp.default.configurationProvider": "ms-vscode.cmake-tools"
}

29
.vscode/tasks.json vendored
View file

@ -1,29 +0,0 @@
{
"version": "2.0.0",
"tasks": [
{
"type": "cmake",
"label": "Configure",
"command": "configure",
"problemMatcher": [
"$gcc"
],
"group": "build"
},
{
"type": "cmake",
"label": "Build",
"command": "build",
"problemMatcher": [
"$gcc"
],
"group": {
"kind": "build",
"isDefault": true
},
"dependsOn": [
"Configure"
]
}
]
}

16
Android.bp
View file

@ -41,7 +41,6 @@ cc_library_static {
cflags: [
"-D_LIB",
"-fPIC",
"-DENABLE_MAIN_REXT_PROFILE",
"-O3",
"-DANDROID",
@ -51,6 +50,8 @@ cc_library_static {
// common/x86/ihevc_sao_ssse3_intr.c: implicit conversion from
// 'int' to 'char' changes value from 128 to -128
"-Wno-error=constant-conversion",
// #KEEP_THREAD_ACTIVE is experimental
"-UKEEP_THREADS_ACTIVE",
],
export_include_dirs: [
@ -65,7 +66,6 @@ cc_library_static {
"common/ihevc_chroma_iquant_itrans_recon.c",
"common/ihevc_chroma_iquant_recon.c",
"common/ihevc_chroma_itrans_recon.c",
"common/ihevc_chroma_itrans_recon_32x32.c",
"common/ihevc_chroma_itrans_recon_16x16.c",
"common/ihevc_chroma_itrans_recon_8x8.c",
"common/ihevc_chroma_recon.c",
@ -79,7 +79,6 @@ cc_library_static {
"common/ihevc_iquant_itrans_recon.c",
"common/ihevc_iquant_recon.c",
"common/ihevc_itrans.c",
"common/ihevc_itrans_res.c",
"common/ihevc_itrans_recon.c",
"common/ihevc_itrans_recon_16x16.c",
"common/ihevc_itrans_recon_32x32.c",
@ -205,6 +204,7 @@ cc_library_static {
"decoder/arm/ihevcd_function_selector_noneon.c",
"decoder/arm64/ihevcd_fmt_conv_420sp_to_420p.s",
"decoder/arm64/ihevcd_fmt_conv_420sp_to_420sp.s",
"decoder/arm64/ihevcd_fmt_conv_420sp_to_rgba8888.s",
"decoder/arm64/ihevcd_function_selector_av8.c",
"decoder/arm64/ihevcd_itrans_recon_dc_chroma.s",
"decoder/arm64/ihevcd_itrans_recon_dc_luma.s",
@ -284,6 +284,7 @@ cc_library_static {
"common/arm/ihevc_weighted_pred_uni.s",
"decoder/arm/ihevcd_fmt_conv_420sp_to_420p.s",
"decoder/arm/ihevcd_fmt_conv_420sp_to_420sp.s",
"decoder/arm/ihevcd_fmt_conv_420sp_to_rgba8888.s",
"decoder/arm/ihevcd_function_selector_a9q.c",
"decoder/arm/ihevcd_itrans_recon_dc_chroma.s",
"decoder/arm/ihevcd_itrans_recon_dc_luma.s",
@ -418,17 +419,19 @@ cc_library_static {
min_sdk_version: "29",
}
cc_binary {
cc_test {
name: "hevcdec",
host_supported: true,
cflags: [
"-DPROFILE_ENABLE",
"-DDISABLE_SEI",
"-DARM",
"-fPIC",
"-DMD5_DISABLE",
"-Wall",
"-Werror",
],
gtest: false,
srcs: ["test/decoder/main.c"],
static_libs: ["libhevcdec"],
target: {
@ -461,7 +464,6 @@ cc_library_static {
"common/ihevc_cabac_tables.c",
"common/ihevc_chroma_intra_pred_filters.c",
"common/ihevc_chroma_itrans_recon.c",
"common/ihevc_chroma_itrans_recon_32x32.c",
"common/ihevc_chroma_itrans_recon_16x16.c",
"common/ihevc_chroma_itrans_recon_8x8.c",
"common/ihevc_common_tables.c",
@ -804,15 +806,17 @@ cc_library_static {
min_sdk_version: "29",
}
cc_binary {
cc_test {
name: "hevcenc",
host_supported: true,
cflags: [
"-DARM",
"-DDISABLE_SEI",
"-fPIC",
"-Wall",
"-Werror",
],
gtest: false,
srcs: ["test/encoder/main.c"],
static_libs: ["libhevcenc"],
sanitize: {

16
CMakeLists.txt
View file

@ -1,17 +1,6 @@
cmake_minimum_required(VERSION 3.9.1)
project(libhevc C CXX)
if(NOT DEFINED SYSTEM_NAME)
set(SYSTEM_NAME ${CMAKE_HOST_SYSTEM_NAME})
endif()
if(NOT DEFINED SYSTEM_PROCESSOR)
set(SYSTEM_PROCESSOR ${CMAKE_HOST_SYSTEM_PROCESSOR})
endif()
if(NOT "${SYSTEM_NAME}" STREQUAL "Darwin")
enable_language(ASM)
endif()
enable_language(ASM)
set(HEVC_ROOT "${CMAKE_CURRENT_SOURCE_DIR}")
set(HEVC_CONFIG_DIR "${CMAKE_CURRENT_BINARY_DIR}")
@ -42,7 +31,6 @@ endif()
include("${HEVC_ROOT}/cmake/utils.cmake")
libhevc_add_compile_options()
libhevc_add_gtest()
libhevc_add_definitions()
libhevc_set_link_libraries()
@ -55,5 +43,3 @@ include("${HEVC_ROOT}/test/encoder/hevcenc.cmake")
include("${HEVC_ROOT}/fuzzer/hevc_dec_fuzzer.cmake")
include("${HEVC_ROOT}/fuzzer/hevc_enc_fuzzer.cmake")
include("${HEVC_ROOT}/tests/common/common.cmake")

28
README.md
View file

@ -7,8 +7,6 @@ Supports:
- aarch32/aarch64 on Linux.
- aarch32/aarch64 on Android.
- x86_32/x86_64 on Linux.
- aarch64 on Mac.
- x86_64 on Mac.
## Native Builds
Use the following commands for building on the target machine
@ -53,29 +51,3 @@ $ make
$ cmake .. -DCMAKE_TOOLCHAIN_FILE=../cmake/toolchains/aarch32_toolchain.cmake
$ make
```
### Building for android
NOTE: This assumes that you are building on a machine that has
[Android NDK](https://developer.android.com/ndk/downloads).
```
$ cd external/libhevc
$ mkdir build
$ cd build
```
#### Armv7 (32-bit)
cmake -DCMAKE_TOOLCHAIN_FILE=../cmake/toolchains/android_toolchain.cmake\
-DHEVC_ANDROID_NDK_PATH=/opt/android-ndk-r26d/\
-DANDROID_ABI=armeabi-v7a\
-DANDROID_PLATFORM=android-23 ../
make
#### Armv8 (64-bit)
cmake -DCMAKE_TOOLCHAIN_FILE=../cmake/toolchains/android_toolchain.cmake\
-DHEVC_ANDROID_NDK_PATH=/opt/android-ndk-r26d/\
-DANDROID_ABI=arm64-v8a\
-DANDROID_PLATFORM=android-23 ../
make

7
cmake/toolchains/aarch32_toolchain.cmake
View file

@ -1,10 +1,7 @@
set(SYSTEM_NAME Linux)
set(SYSTEM_PROCESSOR aarch32)
set(CMAKE_SYSTEM_NAME Linux)
set(CMAKE_SYSTEM_PROCESSOR aarch32)
# Modify these variables with paths to appropriate compilers that can produce
# armv7 targets
set(CMAKE_C_COMPILER arm-linux-gnueabihf-gcc)
set(CMAKE_CXX_COMPILER arm-linux-gnueabihf-g++)
# Build all binaries as static, so that they can be run using qemu
set(CMAKE_EXE_LINKER_FLAGS "-static")

7
cmake/toolchains/aarch64_toolchain.cmake
View file

@ -1,5 +1,5 @@
set(SYSTEM_NAME Linux)
set(SYSTEM_PROCESSOR aarch64)
set(CMAKE_SYSTEM_NAME Linux)
set(CMAKE_SYSTEM_PROCESSOR aarch64)
# Modify these variables with paths to appropriate compilers that can produce
# armv8 targets
@ -11,6 +11,3 @@ set(CMAKE_C_COMPILER_AR
set(CMAKE_CXX_COMPILER_AR
aarch64-linux-gnu-gcc-ar
CACHE FILEPATH "Archiver")
# Build all binaries as static, so that they can be run using qemu
set(CMAKE_EXE_LINKER_FLAGS "-static")

34
cmake/toolchains/android_toolchain.cmake
View file

@ -1,34 +0,0 @@
set(SYSTEM_NAME Android)
set(CMAKE_SYSTEM_NAME Android)
if(NOT ANDROID_PLATFORM)
set(ANDROID_PLATFORM android-23)
endif()
# Choose target architecture with:
# -DANDROID_ABI={armeabi-v7a, arm64-v8a, x86, x86_64}
if(NOT ANDROID_ABI)
set(ANDROID_ABI arm64-v8a)
endif()
if(ANDROID_ABI MATCHES "^armeabi")
set(SYSTEM_PROCESSOR aarch32)
else()
set(SYSTEM_PROCESSOR aarch64)
endif()
# Toolchain files don't have access to cached variables:
# https://gitlab.kitware.com/cmake/cmake/issues/16170. Set an intermediate
# environment variable when loaded the first time.
if(HEVC_ANDROID_NDK_PATH)
set(ENV{HEVC_ANDROID_NDK_PATH} "${HEVC_ANDROID_NDK_PATH}")
else()
set(HEVC_ANDROID_NDK_PATH "$ENV{HEVC_ANDROID_NDK_PATH}")
endif()
if(NOT HEVC_ANDROID_NDK_PATH)
message(FATAL_ERROR "HEVC_ANDROID_NDK_PATH not set.")
return()
endif()
include("${HEVC_ANDROID_NDK_PATH}/build/cmake/android.toolchain.cmake")

46
cmake/utils.cmake
View file

@ -2,9 +2,9 @@ include(CheckCXXCompilerFlag)
# Adds compiler options for all targets
function(libhevc_add_compile_options)
if("${SYSTEM_PROCESSOR}" STREQUAL "aarch64" OR "${SYSTEM_PROCESSOR}" STREQUAL "arm64")
if(${CMAKE_SYSTEM_PROCESSOR} STREQUAL "aarch64")
add_compile_options(-march=armv8-a)
elseif("${SYSTEM_PROCESSOR}" STREQUAL "aarch32")
elseif(${CMAKE_SYSTEM_PROCESSOR} STREQUAL "aarch32")
add_compile_options(-march=armv7-a -mfpu=neon)
else()
add_compile_options(-msse4.2 -mno-avx)
@ -32,15 +32,9 @@ endfunction()
# Adds defintions for all targets
function(libhevc_add_definitions)
if("${SYSTEM_NAME}" STREQUAL "Darwin")
if("${SYSTEM_PROCESSOR}" STREQUAL "arm64")
add_definitions(-DARMV8 -DDARWIN -DDEFAULT_ARCH=D_ARCH_ARMV8_GENERIC)
else()
add_definitions(-DX86 -DDARWIN -DDISABLE_AVX2 -DDEFAULT_ARCH=D_ARCH_X86_GENERIC)
endif()
elseif("${SYSTEM_PROCESSOR}" STREQUAL "aarch64")
if(${CMAKE_SYSTEM_PROCESSOR} STREQUAL "aarch64")
add_definitions(-DARMV8 -DDEFAULT_ARCH=D_ARCH_ARMV8_GENERIC -DENABLE_NEON)
elseif("${SYSTEM_PROCESSOR}" STREQUAL "aarch32")
elseif(${CMAKE_SYSTEM_PROCESSOR} STREQUAL "aarch32")
add_definitions(-DARMV7 -DDEFAULT_ARCH=D_ARCH_ARM_A9Q -DENABLE_NEON
-DDISABLE_NEONINTR)
else()
@ -118,35 +112,3 @@ endfunction()
function(libhevc_add_fuzzer NAME LIB)
libhevc_add_executable(${NAME} ${LIB} FUZZER 1 ${ARGV})
endfunction()
# Adds GoogleTest and Threads dependency
function(libhevc_add_gtest)
include(FetchContent)
FetchContent_Declare(
googletest
URL https://github.com/google/googletest/archive/03597a01ee50ed33e9dfd640b249b4be3799d395.zip
)
# For Windows: Prevent overriding the parent project's compiler/linker settings
set(gtest_force_shared_crt ON CACHE BOOL "" FORCE)
FetchContent_MakeAvailable(googletest)
endfunction()
# cmake-format: off
# Adds a target for a gtest executable
#
# Arguments:
# NAME: Name of the executable
#
# Optional Arguments:
# SOURCES: Additional source files
# cmake-format: on
function(libhevc_add_gtest_executable NAME)
set(multi_value_args SOURCES)
cmake_parse_arguments(ARG "" "" "${multi_value_args}" ${ARGN})
libhevc_add_executable(
${NAME} libhevcdec
SOURCES ${HEVC_ROOT}/tests/common/func_selector.cc
${HEVC_ROOT}/tests/common/tests_common.cc ${ARG_SOURCES}
LIBS GTest::gtest_main)
endfunction()

15
common/arm/ihevc_intra_pred_filters_neon_intr.c
View file

@ -438,11 +438,6 @@ void ihevc_intra_pred_luma_ref_substitution_neonintr(UWORD8 *pu1_top_left,
* @param[in] mode
* integer intraprediction mode
*
* @param[in] intra_smoothing_flags
* integer bit 3 indicates if intra smoothing is enabled/disabled
* unconditionally. this is applicable to frext profiles only
* bit 0 indicates strong intra smoothing enabled/disabled
*
* @returns
*
* @remarks
@ -456,7 +451,7 @@ void ihevc_intra_pred_ref_filtering_neonintr(UWORD8 *pu1_src,
WORD32 nt,
UWORD8 *pu1_dst,
WORD32 mode,
WORD32 intra_smoothing_flags)
WORD32 strong_intra_smoothing_enable_flag)
{
WORD32 filter_flag;
WORD32 i = 0;
@ -480,12 +475,10 @@ void ihevc_intra_pred_ref_filtering_neonintr(UWORD8 *pu1_src,
WORD32 abs_cond_left_flag = 0;
WORD32 abs_cond_top_flag = 0;
WORD32 dc_val = 1 << (BIT_DEPTH - 5);
WORD32 intra_smoothing_disabled = (intra_smoothing_flags >> 3);
WORD32 strong_intra_smoothing_enable_flag = intra_smoothing_flags & 1;
shift_res = vdup_n_u8(0);
filter_flag = intra_smoothing_disabled ?
0 : (gau1_intra_pred_ref_filter[mode] & (1 << (CTZ(nt) - 2)));
filter_flag = gau1_intra_pred_ref_filter[mode] & (1 << (CTZ(nt) - 2));
if(0 == filter_flag)
{
if(pu1_src == pu1_dst)

47
common/arm/ihevc_intra_ref_substitution_a9q.c
View file

@ -103,8 +103,7 @@ void ihevc_intra_pred_chroma_ref_substitution_a9q(UWORD8 *pu1_top_left,
WORD32 nt,
WORD32 nbr_flags,
UWORD8 *pu1_dst,
WORD32 dst_strd,
WORD32 chroma_format_idc)
WORD32 dst_strd)
{
UWORD8 pu1_ref_u, pu1_ref_v;
WORD32 dc_val, i, j;
@ -181,7 +180,7 @@ void ihevc_intra_pred_chroma_ref_substitution_a9q(UWORD8 *pu1_top_left,
// U-V interleaved Top-top right samples
}
if(nt == 4 || (nt == 8 && chroma_format_idc == CHROMA_FMT_IDC_YUV444))
if(nt == 4)
{
/* 1 bit extraction for all the neighboring blocks */
tp_left = (nbr_flags & 0x10000) >> 16;
@ -249,9 +248,8 @@ void ihevc_intra_pred_chroma_ref_substitution_a9q(UWORD8 *pu1_top_left,
}
}
else if(nt == 8 || (nt == 16 && chroma_format_idc == CHROMA_FMT_IDC_YUV444))
else if(nt == 8)
{
WORD32 sub_sample = chroma_format_idc == CHROMA_FMT_IDC_YUV444 ? 2 : 1;
WORD32 nbr_flags_temp = 0;
nbr_flags_temp = ((nbr_flags & 0xC) >> 2) + ((nbr_flags & 0xC0) >> 4)
+ ((nbr_flags & 0x300) >> 4)
@ -261,16 +259,16 @@ void ihevc_intra_pred_chroma_ref_substitution_a9q(UWORD8 *pu1_top_left,
/* compute trailing zeors based on nbr_flag for substitution process of below left see section .*/
/* as each bit in nbr flags corresponds to 8 pels for bot_left, left, top and topright but 1 pel for topleft */
{
nbr_id_from_bl = look_up_trailing_zeros(nbr_flags_temp & 0XF) * (4 * sub_sample); /* for bottom left and left */
if(nbr_id_from_bl == 32 * sub_sample)
nbr_id_from_bl = 16 * sub_sample;
if(nbr_id_from_bl == 16 * sub_sample)
nbr_id_from_bl = look_up_trailing_zeros(nbr_flags_temp & 0XF) * 4; /* for bottom left and left */
if(nbr_id_from_bl == 32)
nbr_id_from_bl = 16;
if(nbr_id_from_bl == 16)
{
/* for top left : 1 pel per nbr bit */
if(!((nbr_flags_temp >> 8) & 0x1))
{
nbr_id_from_bl++;
nbr_id_from_bl += look_up_trailing_zeros((nbr_flags_temp >> 4) & 0xF) * 4 * sub_sample; /* top and top right; 8 pels per nbr bit */
nbr_id_from_bl += look_up_trailing_zeros((nbr_flags_temp >> 4) & 0xF) * 4; /* top and top right; 8 pels per nbr bit */
}
}
@ -289,14 +287,14 @@ void ihevc_intra_pred_chroma_ref_substitution_a9q(UWORD8 *pu1_top_left,
}
/* for the loop of 4*Nt+1 pixels (excluding pixels computed from reverse substitution) */
while(nbr_id_from_bl < ((T8C_4NT * sub_sample)+1))
while(nbr_id_from_bl < ((T8C_4NT)+1))
{
/* To Obtain the next unavailable idx flag after reverse neighbor substitution */
/* Divide by 8 to obtain the original index */
frwd_nbr_flag = (nbr_id_from_bl >> (chroma_format_idc == CHROMA_FMT_IDC_YUV444 ? 3 : 2)); /*+ (nbr_id_from_bl & 0x1);*/
frwd_nbr_flag = (nbr_id_from_bl >> 2); /*+ (nbr_id_from_bl & 0x1);*/
/* The Top-left flag is at the last bit location of nbr_flags*/
if(nbr_id_from_bl == (T8C_4NT * sub_sample / 2))
if(nbr_id_from_bl == (T8C_4NT / 2))
{
get_bits = GET_BIT(nbr_flags_temp, 8);
@ -315,23 +313,22 @@ void ihevc_intra_pred_chroma_ref_substitution_a9q(UWORD8 *pu1_top_left,
UWORD16 *pu2_dst;
/* 8 pel substitution (other than TL) */
pu2_dst = (UWORD16 *)&pu1_dst[(2 * nbr_id_from_bl) - 2];
ihevc_memset_16bit_a9q((UWORD16 *)(pu1_dst + (2 * nbr_id_from_bl)), pu2_dst[0], 4 * sub_sample);
ihevc_memset_16bit_a9q((UWORD16 *)(pu1_dst + (2 * nbr_id_from_bl)), pu2_dst[0], 4);
}
}
nbr_id_from_bl += (nbr_id_from_bl == (T8C_4NT * sub_sample / 2)) ? 1 : 4 * sub_sample;
nbr_id_from_bl += (nbr_id_from_bl == (T8C_4NT / 2)) ? 1 : 4;
}
}
else if(nt == 16 || (nt == 32 && chroma_format_idc == CHROMA_FMT_IDC_YUV444))
else if(nt == 16)
{
WORD32 sub_sample = chroma_format_idc == CHROMA_FMT_IDC_YUV444 ? 2 : 1;
/* compute trailing ones based on mbr_flag for substitution process of below left see section .*/
/* as each bit in nbr flags corresponds to 4 pels for bot_left, left, top and topright but 1 pel for topleft */
{
nbr_id_from_bl = look_up_trailing_zeros((nbr_flags & 0XFF)) * 4 * sub_sample; /* for bottom left and left */
nbr_id_from_bl = look_up_trailing_zeros((nbr_flags & 0XFF)) * 4; /* for bottom left and left */
if(nbr_id_from_bl == 32 * sub_sample)
if(nbr_id_from_bl == 32)
{
/* for top left : 1 pel per nbr bit */
if(!((nbr_flags >> 16) & 0x1))
@ -339,7 +336,7 @@ void ihevc_intra_pred_chroma_ref_substitution_a9q(UWORD8 *pu1_top_left,
/* top left not available */
nbr_id_from_bl++;
/* top and top right; 4 pels per nbr bit */
nbr_id_from_bl += look_up_trailing_zeros((nbr_flags >> 8) & 0xFF) * 4 * sub_sample;
nbr_id_from_bl += look_up_trailing_zeros((nbr_flags >> 8) & 0xFF) * 4;
}
}
/* Reverse Substitution Process*/
@ -357,14 +354,14 @@ void ihevc_intra_pred_chroma_ref_substitution_a9q(UWORD8 *pu1_top_left,
}
/* for the loop of 4*Nt+1 pixels (excluding pixels computed from reverse substitution) */
while(nbr_id_from_bl < ((T16C_4NT * sub_sample)+1))
while(nbr_id_from_bl < ((T16C_4NT)+1))
{
/* To Obtain the next unavailable idx flag after reverse neighbor substitution */
/* Devide by 4 to obtain the original index */
frwd_nbr_flag = (nbr_id_from_bl >> (chroma_format_idc == CHROMA_FMT_IDC_YUV444 ? 3 : 2)); /*+ (nbr_id_from_bl & 0x1);*/
frwd_nbr_flag = (nbr_id_from_bl >> 2); /*+ (nbr_id_from_bl & 0x1);*/
/* The Top-left flag is at the last bit location of nbr_flags*/
if(nbr_id_from_bl == (T16C_4NT * sub_sample / 2))
if(nbr_id_from_bl == (T16C_4NT / 2))
{
get_bits = GET_BIT(nbr_flags, 16);
/* only pel substitution for TL */
@ -382,11 +379,11 @@ void ihevc_intra_pred_chroma_ref_substitution_a9q(UWORD8 *pu1_top_left,
UWORD16 *pu2_dst;
/* 4 pel substitution (other than TL) */
pu2_dst = (UWORD16 *)&pu1_dst[(2 * nbr_id_from_bl) - 2];
ihevc_memset_16bit_a9q((UWORD16 *)(pu1_dst + (2 * nbr_id_from_bl)), pu2_dst[0], 4 * sub_sample);
ihevc_memset_16bit_a9q((UWORD16 *)(pu1_dst + (2 * nbr_id_from_bl)), pu2_dst[0], 4);
}
}
nbr_id_from_bl += (nbr_id_from_bl == (T16C_4NT * sub_sample / 2)) ? 1 : 4 * sub_sample;
nbr_id_from_bl += (nbr_id_from_bl == (T16C_4NT / 2)) ? 1 : 4;
}
}
}

6
common/arm64/ihevc_deblk_chroma_horz.s
View file

@ -45,8 +45,9 @@
// WORD32 filter_flag_q)
//
.include "ihevc_neon_macros.s"
.text
.align 4
.include "ihevc_neon_macros.s"
@ -56,7 +57,7 @@
.type ihevc_deblk_chroma_horz_av8, %function
ENTRY ihevc_deblk_chroma_horz_av8
ihevc_deblk_chroma_horz_av8:
sxtw x4,w4
sxtw x5,w5
sxtw x6,w6
@ -165,7 +166,6 @@ l1.3528:
l1.3540:
ldp x19, x20,[sp],#16
pop_v_regs
EXIT_FUNC
ret

6
common/arm64/ihevc_deblk_chroma_vert.s
View file

@ -46,8 +46,9 @@
// WORD32 filter_flag_p,
// WORD32 filter_flag_q)
.include "ihevc_neon_macros.s"
.text
.align 4
.include "ihevc_neon_macros.s"
@ -57,7 +58,7 @@
.type ihevc_deblk_chroma_vert_av8, %function
ENTRY ihevc_deblk_chroma_vert_av8
ihevc_deblk_chroma_vert_av8:
sxtw x4,w4
sxtw x5,w5
sxtw x6,w6
@ -204,7 +205,6 @@ l1.3204:
l1.3228:
ldp x19, x20,[sp],#16
pop_v_regs
EXIT_FUNC
ret

6
common/arm64/ihevc_deblk_luma_horz.s
View file

@ -36,8 +36,8 @@
//*
//*******************************************************************************/
.include "ihevc_neon_macros.s"
.text
.align 4
.extern gai4_ihevc_tc_table
@ -46,7 +46,7 @@
.type ihevc_deblk_luma_horz_av8, %function
ENTRY ihevc_deblk_luma_horz_av8
ihevc_deblk_luma_horz_av8:
// stmfd sp!, {x3-x12,x14}
sxtw x5,w5
sxtw x6,w6
@ -434,7 +434,6 @@ l1.2404:
ldp d10,d11,[sp],#16
ldp d8,d9,[sp],#16 // Loading d9 using { ldr d9,[sp]; add sp,sp,#8 } is giving bus error.
// d8 is used as dummy register and loaded along with d9 using ldp. d8 is not used in the function.
EXIT_FUNC
ret
// x4=flag p
@ -585,7 +584,6 @@ l1.2852:
ldp d10,d11,[sp],#16
ldp d8,d9,[sp],#16 // Loading d9 using { ldr d9,[sp]; add sp,sp,#8 } is giving bus error.
// d8 is used as dummy register and loaded along with d9 using ldp. d8 is not used in the function.
EXIT_FUNC
ret

6
common/arm64/ihevc_deblk_luma_vert.s
View file

@ -37,8 +37,8 @@
//*
//*******************************************************************************/
.include "ihevc_neon_macros.s"
.text
.align 4
@ -49,7 +49,7 @@
.type ihevc_deblk_luma_vert_av8, %function
ENTRY ihevc_deblk_luma_vert_av8
ihevc_deblk_luma_vert_av8:
sxtw x5,w5
sxtw x6,w6
@ -450,7 +450,6 @@ l1.964:
ldp d12,d13,[sp],#16
ldp d10,d11,[sp],#16
ldp d8,d9,[sp],#16
EXIT_FUNC
ret
l1.968:
@ -631,7 +630,6 @@ l1.1412:
ldp d12,d13,[sp],#16
ldp d10,d11,[sp],#16
ldp d8,d9,[sp],#16
EXIT_FUNC
ret

10
common/arm64/ihevc_inter_pred_chroma_copy.s
View file

@ -91,14 +91,14 @@
//x5 => ht
//x6 => wd
.include "ihevc_neon_macros.s"
.text
.align 4
.globl ihevc_inter_pred_chroma_copy_av8
.type ihevc_inter_pred_chroma_copy_av8, %function
ENTRY ihevc_inter_pred_chroma_copy_av8
ihevc_inter_pred_chroma_copy_av8:
LSL x12,x6,#1 //wd << 1
CMP x5,#0 //checks ht == 0
@ -142,8 +142,7 @@ END_INNER_LOOP_WD_4:
BGT OUTER_LOOP_WD_4_HT_2
END_LOOPS:
EXIT_FUNC
ret
RET
OUTER_LOOP_WD_4_HT_2:
SUBS x4,x12,#0 //checks wd == 0
@ -252,7 +251,6 @@ INNER_LOOP_WD_16_HT_2:
LD1 {v1.16b},[x7],x2 //vld1_u8(pu1_src_tmp)
ST1 {v1.16b},[x6],x3 //vst1_u8(pu1_dst_tmp, tmp_src)
EXIT_FUNC
ret
RET

8
common/arm64/ihevc_inter_pred_chroma_copy_w16out.s
View file

@ -92,14 +92,16 @@
//x5 => ht
//x6 => wd
.include "ihevc_neon_macros.s"
.text
.align 4
.include "ihevc_neon_macros.s"
.globl ihevc_inter_pred_chroma_copy_w16out_av8
.type ihevc_inter_pred_chroma_copy_w16out_av8, %function
ENTRY ihevc_inter_pred_chroma_copy_w16out_av8
ihevc_inter_pred_chroma_copy_w16out_av8:
// stmfd sp!, {x4-x12, x14} //stack stores the values of the arguments
@ -171,7 +173,6 @@ end_loops:
// ldmfd sp!,{x4-x12,x15} //reload the registers from sp
ldp x19, x20,[sp],#16
EXIT_FUNC
ret
@ -338,7 +339,6 @@ core_loop_wd_8_ht_2:
// ldmfd sp!,{x4-x12,x15} //reload the registers from sp
ldp x19, x20,[sp],#16
EXIT_FUNC
ret

6
common/arm64/ihevc_inter_pred_chroma_horz.s
View file

@ -93,15 +93,16 @@
//x2 => src_strd
//x3 => dst_strd
.text
.align 4
.include "ihevc_neon_macros.s"
.text
.globl ihevc_inter_pred_chroma_horz_av8
.type ihevc_inter_pred_chroma_horz_av8, %function
ENTRY ihevc_inter_pred_chroma_horz_av8
ihevc_inter_pred_chroma_horz_av8:
// stmfd sp!, {x4-x12, x14} //stack stores the values of the arguments
@ -768,7 +769,6 @@ end_loops:
ldp d13,d14,[sp],#16
ldp d11,d12,[sp],#16
ldp d9,d10,[sp],#16
EXIT_FUNC
ret

6
common/arm64/ihevc_inter_pred_chroma_horz_w16out.s
View file

@ -91,16 +91,17 @@
//x3 => dst_strd
.text
.align 4
.include "ihevc_neon_macros.s"
.text
.globl ihevc_inter_pred_chroma_horz_w16out_av8
.type ihevc_inter_pred_chroma_horz_w16out_av8, %function
ENTRY ihevc_inter_pred_chroma_horz_w16out_av8
ihevc_inter_pred_chroma_horz_w16out_av8:
// stmfd sp!, {x4-x12, x14} //stack stores the values of the arguments
@ -793,7 +794,6 @@ end_loops:
ldp d14,d15,[sp],#16
ldp d12,d13,[sp],#16
ldp d10,d11,[sp],#16
EXIT_FUNC
ret

6
common/arm64/ihevc_inter_pred_chroma_vert.s
View file

@ -92,15 +92,16 @@
//x1 => *pi2_dst
//x2 => src_strd
//x3 => dst_strd
.text
.align 4
.include "ihevc_neon_macros.s"
.text
.globl ihevc_inter_pred_chroma_vert_av8
.type ihevc_inter_pred_chroma_vert_av8, %function
ENTRY ihevc_inter_pred_chroma_vert_av8
ihevc_inter_pred_chroma_vert_av8:
// stmfd sp!,{x4-x12,x14} //stack stores the values of the arguments
@ -398,7 +399,6 @@ end_loops:
// ldmfd sp!,{x4-x12,x15} //reload the registers from sp
ldp x19, x20,[sp],#16
EXIT_FUNC
ret

7
common/arm64/ihevc_inter_pred_chroma_vert_w16inp.s
View file

@ -92,14 +92,16 @@
//x2 => src_strd
//x3 => dst_strd
.include "ihevc_neon_macros.s"
.text
.align 4
.include "ihevc_neon_macros.s"
.globl ihevc_inter_pred_chroma_vert_w16inp_av8
.type ihevc_inter_pred_chroma_vert_w16inp_av8, %function
ENTRY ihevc_inter_pred_chroma_vert_w16inp_av8
ihevc_inter_pred_chroma_vert_w16inp_av8:
// stmfd sp!, {x4-x12, x14} //stack stores the values of the arguments
@ -347,7 +349,6 @@ end_loops:
// ldmfd sp!,{x4-x12,x15} //reload the registers from sp
ldp x19, x20,[sp],#16
EXIT_FUNC
ret

6
common/arm64/ihevc_inter_pred_chroma_vert_w16inp_w16out.s
View file

@ -92,15 +92,16 @@
//x1 => *pi2_dst
//x2 => src_strd
//x3 => dst_strd
.text
.align 4
.include "ihevc_neon_macros.s"
.text
.globl ihevc_inter_pred_chroma_vert_w16inp_w16out_av8
.type ihevc_inter_pred_chroma_vert_w16inp_w16out_av8, %function
ENTRY ihevc_inter_pred_chroma_vert_w16inp_w16out_av8
ihevc_inter_pred_chroma_vert_w16inp_w16out_av8:
// stmfd sp!, {x4-x12, x14} //stack stores the values of the arguments
@ -335,7 +336,6 @@ end_loops:
// ldmfd sp!,{x4-x12,x15} //reload the registers from sp
ldp x19, x20,[sp],#16
EXIT_FUNC
ret

7
common/arm64/ihevc_inter_pred_chroma_vert_w16out.s
View file

@ -93,14 +93,16 @@
//x2 => src_strd
//x3 => dst_strd
.include "ihevc_neon_macros.s"
.text
.align 4
.include "ihevc_neon_macros.s"
.globl ihevc_inter_pred_chroma_vert_w16out_av8
.type ihevc_inter_pred_chroma_vert_w16out_av8, %function
ENTRY ihevc_inter_pred_chroma_vert_w16out_av8
ihevc_inter_pred_chroma_vert_w16out_av8:
// stmfd sp!,{x4-x12,x14} //stack stores the values of the arguments
@ -384,7 +386,6 @@ end_loops:
// ldmfd sp!,{x4-x12,x15} //reload the registers from sp
ldp x19, x20,[sp],#16
EXIT_FUNC
ret

8
common/arm64/ihevc_inter_pred_filters_luma_horz.s
View file

@ -103,15 +103,16 @@
// x5 => ht
// x6 => wd
.text
.align 4
.include "ihevc_neon_macros.s"
.text
.globl ihevc_inter_pred_luma_horz_av8
.type ihevc_inter_pred_luma_horz_av8, %function
ENTRY ihevc_inter_pred_luma_horz_av8
ihevc_inter_pred_luma_horz_av8:
// stmfd sp!, {x4-x12, x14} //stack stores the values of the arguments
push_v_regs
@ -285,7 +286,6 @@ end_loops:
// ldmfd sp!,{x4-x12,x15} //reload the registers from sp
ldp x19, x20,[sp], #16
pop_v_regs
EXIT_FUNC
ret
@ -481,7 +481,6 @@ end_loops1:
// ldmfd sp!,{x4-x12,x15} //reload the registers from sp
ldp x19, x20,[sp], #16
pop_v_regs
EXIT_FUNC
ret
@ -596,7 +595,6 @@ end_inner_loop_4:
// ldmfd sp!,{x4-x12,x15} //reload the registers from sp
ldp x19, x20,[sp], #16
pop_v_regs
EXIT_FUNC
ret

7
common/arm64/ihevc_inter_pred_filters_luma_vert.s
View file

@ -103,15 +103,16 @@
// x12 => *pi1_coeff
// x5 => ht
// x3 => wd
.text
.align 4
.include "ihevc_neon_macros.s"
.text
.globl ihevc_inter_pred_luma_vert_av8
.type ihevc_inter_pred_luma_vert_av8, %function
ENTRY ihevc_inter_pred_luma_vert_av8
ihevc_inter_pred_luma_vert_av8:
// stmfd sp!, {x4-x12, x14} //stack stores the values of the arguments
@ -427,7 +428,6 @@ end_loops:
bne lbl409
ldp x19, x20,[sp], #16
EXIT_FUNC
ret
lbl409:
mov x5, #4
@ -518,6 +518,5 @@ end_inner_loop_wd_4:
// ldmfd sp!, {x4-x12, x15} //reload the registers from sp
ldp x19, x20,[sp], #16
EXIT_FUNC
ret

6
common/arm64/ihevc_inter_pred_filters_luma_vert_w16inp.s
View file

@ -94,15 +94,16 @@
// word32 ht,
// word32 wd )
.text
.align 4
.include "ihevc_neon_macros.s"
.text
.globl ihevc_inter_pred_luma_vert_w16inp_av8
.type ihevc_inter_pred_luma_vert_w16inp_av8, %function
ENTRY ihevc_inter_pred_luma_vert_w16inp_av8
ihevc_inter_pred_luma_vert_w16inp_av8:
// stmfd sp!, {x4-x12, x14} //stack stores the values of the arguments
@ -383,7 +384,6 @@ end_loops:
// ldmfd sp!,{x4-x12,x15} //reload the registers from sp
ldp x19, x20,[sp], #16
EXIT_FUNC
ret

5
common/arm64/ihevc_inter_pred_filters_luma_vert_w16out.s
View file

@ -62,13 +62,12 @@
.include "ihevc_neon_macros.s"
.text
.globl ihevc_inter_pred_luma_vert_w16out_av8
.type ihevc_inter_pred_luma_vert_w16out_av8, %function
ENTRY ihevc_inter_pred_luma_vert_w16out_av8
ihevc_inter_pred_luma_vert_w16out_av8:
// stmfd sp!, {x4-x12, x14} //stack stores the values of the arguments
@ -379,7 +378,6 @@ end_loops_16out:
bne lbl355
ldp x19, x20,[sp], #16
EXIT_FUNC
ret
lbl355:
mov x5, #4
@ -473,7 +471,6 @@ end_inner_loop_wd_4_16out:
// ldmfd sp!, {x4-x12, x15} //reload the registers from sp
ldp x19, x20,[sp], #16
EXIT_FUNC
ret

9
common/arm64/ihevc_inter_pred_luma_copy.s
View file

@ -71,14 +71,16 @@
// x11 => ht
// x16 => wd
.include "ihevc_neon_macros.s"
.text
.align 4
.include "ihevc_neon_macros.s"
.globl ihevc_inter_pred_luma_copy_av8
.type ihevc_inter_pred_luma_copy_av8, %function
ENTRY ihevc_inter_pred_luma_copy_av8
ihevc_inter_pred_luma_copy_av8:
// stmfd sp!, {x8-x16, lr} //stack stores the values of the arguments
stp x19,x20,[sp, #-16]!
mov x16,x6 //loads wd
@ -123,7 +125,6 @@ end_loops:
// MRS x20,PMCCFILTR_EL0
sub x0,x20,x19
ldp x19,x20,[sp],#16
EXIT_FUNC
ret
@ -158,7 +159,6 @@ end_inner_loop_wd_8:
// MRS x20,PMCCFILTR_EL0
sub x0,x20,x19
ldp x19,x20,[sp],#16
EXIT_FUNC
ret
core_loop_wd_16:
@ -192,7 +192,6 @@ end_inner_loop_wd_16:
// MRS x20,PMCCFILTR_EL0
sub x0,x20,x19
ldp x19,x20,[sp],#16
EXIT_FUNC
ret

7
common/arm64/ihevc_inter_pred_luma_copy_w16out.s
View file

@ -72,15 +72,16 @@
// x7 => ht
// x12 => wd
.text
.align 4
.include "ihevc_neon_macros.s"
.text
.globl ihevc_inter_pred_luma_copy_w16out_av8
.type ihevc_inter_pred_luma_copy_w16out_av8, %function
ENTRY ihevc_inter_pred_luma_copy_w16out_av8
ihevc_inter_pred_luma_copy_w16out_av8:
// stmfd sp!, {x4-x12, x14} //stack stores the values of the arguments
@ -139,7 +140,6 @@ end_loops:
ldp x19, x20,[sp], #16
EXIT_FUNC
ret
@ -265,7 +265,6 @@ epilog_end:
// ldmfd sp!,{x4-x12,x15} //reload the registers from sp
ldp x19, x20,[sp], #16
EXIT_FUNC
ret

9
common/arm64/ihevc_inter_pred_luma_horz_w16out.s
View file

@ -107,15 +107,16 @@
//x15 - #1
//x16 - src_ptx1
//x19 - loop_counter
.text
.align 4
.include "ihevc_neon_macros.s"
.text
.globl ihevc_inter_pred_luma_horz_w16out_av8
.type ihevc_inter_pred_luma_horz_w16out_av8, %function
ENTRY ihevc_inter_pred_luma_horz_w16out_av8
ihevc_inter_pred_luma_horz_w16out_av8:
// stmfd sp!, {x8-x16, x19} //stack stores the values of the arguments
push_v_regs
@ -304,7 +305,6 @@ height_residue_4:
bne lbl280
ldp x19, x20,[sp], #16
pop_v_regs
EXIT_FUNC
ret
lbl280:
@ -365,7 +365,6 @@ end_inner_loop_height_residue_4:
// ldmfd sp!,{x8-x16,pc} //reload the registers from sp
ldp x19, x20,[sp], #16
pop_v_regs
EXIT_FUNC
ret
outer_loop8_residual:
@ -477,7 +476,6 @@ end_inner_loop_8:
// ldmfd sp!,{x8-x16,pc} //reload the registers from sp
ldp x19, x20,[sp], #16
pop_v_regs
EXIT_FUNC
ret
@ -668,7 +666,6 @@ end_loops1:
// ldmfd sp!,{x8-x16,pc} //reload the registers from sp
ldp x19, x20,[sp], #16
pop_v_regs
EXIT_FUNC
ret

7
common/arm64/ihevc_inter_pred_luma_vert_w16inp_w16out.s
View file

@ -102,14 +102,16 @@
// r5 => ht
// r6 => wd
.include "ihevc_neon_macros.s"
.text
.align 4
.include "ihevc_neon_macros.s"
.globl ihevc_inter_pred_luma_vert_w16inp_w16out_av8
.type ihevc_inter_pred_luma_vert_w16inp_w16out_av8, %function
ENTRY ihevc_inter_pred_luma_vert_w16inp_w16out_av8
ihevc_inter_pred_luma_vert_w16inp_w16out_av8:
//stmfd sp!, {r4-r12, r14} //stack stores the values of the arguments
@ -406,7 +408,6 @@ end_loops:
//ldmfd sp!,{r4-r12,r15} //reload the registers from sp
ldp x19, x20,[sp], #16
EXIT_FUNC
ret

6
common/arm64/ihevc_intra_pred_chroma_dc.s
View file

@ -92,8 +92,9 @@
// mode
// pi1_coeff
.include "ihevc_neon_macros.s"
.text
.align 4
.include "ihevc_neon_macros.s"
@ -101,7 +102,7 @@
.type ihevc_intra_pred_chroma_dc_av8, %function
ENTRY ihevc_intra_pred_chroma_dc_av8
ihevc_intra_pred_chroma_dc_av8:
// stmfd sp!, {x4-x12, x14} //stack stores the values of the arguments
push_v_regs
@ -292,7 +293,6 @@ end_func:
// ldmfd sp!,{x4-x12,x15} //reload the registers from sp
ldp x19, x20,[sp],#16
pop_v_regs
EXIT_FUNC
ret

9
common/arm64/ihevc_intra_pred_chroma_horz.s
View file

@ -84,15 +84,16 @@
//x2 => *pu1_dst
//x3 => dst_strd
.include "ihevc_neon_macros.s"
.text
.align 4
.include "ihevc_neon_macros.s"
.globl ihevc_intra_pred_chroma_horz_av8
.type ihevc_intra_pred_chroma_horz_av8, %function
ENTRY ihevc_intra_pred_chroma_horz_av8
ihevc_intra_pred_chroma_horz_av8:
// stmfd sp!, {x4-x12, x14} //stack stores the values of the arguments
@ -188,7 +189,6 @@ core_loop_16:
// ldmfd sp!,{x4-x12,x15} //reload the registers from sp
ldp x19, x20,[sp],#16
EXIT_FUNC
ret
b endloop
@ -270,7 +270,6 @@ core_loop_8:
// ldmfd sp!,{x4-x12,x15} //reload the registers from sp
ldp x19, x20,[sp],#16
EXIT_FUNC
ret
b endloop
@ -319,7 +318,6 @@ core_loop_4:
// ldmfd sp!,{x4-x12,x15} //reload the registers from sp
ldp x19, x20,[sp],#16
EXIT_FUNC
ret
b endloop
@ -355,7 +353,6 @@ core_loop_4:
// ldmfd sp!,{x4-x12,x15} //reload the registers from sp
ldp x19, x20,[sp],#16
EXIT_FUNC
ret
endloop:

6
common/arm64/ihevc_intra_pred_chroma_mode2.s
View file

@ -92,8 +92,9 @@
// mode
// pi1_coeff
.include "ihevc_neon_macros.s"
.text
.align 4
.include "ihevc_neon_macros.s"
@ -101,7 +102,7 @@
.type ihevc_intra_pred_chroma_mode2_av8, %function
ENTRY ihevc_intra_pred_chroma_mode2_av8
ihevc_intra_pred_chroma_mode2_av8:
// stmfd sp!, {x4-x12, x14} //stack stores the values of the arguments
push_v_regs
@ -302,7 +303,6 @@ end_func:
// ldmfd sp!,{x4-x12,x15} //reload the registers from sp
ldp x19, x20,[sp],#16
pop_v_regs
EXIT_FUNC
ret

6
common/arm64/ihevc_intra_pred_chroma_mode_18_34.s
View file

@ -92,8 +92,9 @@
// mode
// pi1_coeff
.include "ihevc_neon_macros.s"
.text
.align 4
.include "ihevc_neon_macros.s"
@ -101,7 +102,7 @@
.type ihevc_intra_pred_chroma_mode_18_34_av8, %function
ENTRY ihevc_intra_pred_chroma_mode_18_34_av8
ihevc_intra_pred_chroma_mode_18_34_av8:
// stmfd sp!, {x4-x12, x14} //stack stores the values of the arguments
@ -188,7 +189,6 @@ end_func:
// ldmfd sp!,{x4-x12,x15} //reload the registers from sp
ldp x19, x20,[sp],#16
EXIT_FUNC
ret

6
common/arm64/ihevc_intra_pred_chroma_mode_27_to_33.s
View file

@ -81,8 +81,9 @@
// word32 nt,
// word32 mode)
.include "ihevc_neon_macros.s"
.text
.align 4
.include "ihevc_neon_macros.s"
.globl ihevc_intra_pred_chroma_mode_27_to_33_av8
@ -91,7 +92,7 @@
.type ihevc_intra_pred_chroma_mode_27_to_33_av8, %function
ENTRY ihevc_intra_pred_chroma_mode_27_to_33_av8
ihevc_intra_pred_chroma_mode_27_to_33_av8:
// stmfd sp!, {x4-x12, x14} //stack stores the values of the arguments
@ -548,7 +549,6 @@ end_loops:
ldp d14,d15,[sp],#16
ldp d12,d13,[sp],#16
ldp d9,d10,[sp],#16
EXIT_FUNC
ret

7
common/arm64/ihevc_intra_pred_chroma_mode_3_to_9.s
View file

@ -86,8 +86,10 @@
// nt
// mode
.include "ihevc_neon_macros.s"
.text
.align 4
.include "ihevc_neon_macros.s"
@ -99,7 +101,7 @@
.type ihevc_intra_pred_chroma_mode_3_to_9_av8, %function
ENTRY ihevc_intra_pred_chroma_mode_3_to_9_av8
ihevc_intra_pred_chroma_mode_3_to_9_av8:
// stmfd sp!, {x4-x12, x14} //stack stores the values of the arguments
@ -487,7 +489,6 @@ end_func:
ldp d8,d15,[sp],#16 // Loading d15 using { ldr d15,[sp]; add sp,sp,#8 } is giving bus error.
// d8 is used as dummy register and loaded along with d15 using ldp. d8 is not used in the function.
ldp d13,d14,[sp],#16
EXIT_FUNC
ret

6
common/arm64/ihevc_intra_pred_chroma_planar.s
View file

@ -92,8 +92,9 @@
// mode
// pi1_coeff
.include "ihevc_neon_macros.s"
.text
.align 4
.include "ihevc_neon_macros.s"
.globl ihevc_intra_pred_chroma_planar_av8
@ -102,7 +103,7 @@
.type ihevc_intra_pred_chroma_planar_av8, %function
ENTRY ihevc_intra_pred_chroma_planar_av8
ihevc_intra_pred_chroma_planar_av8:
// stmfd sp!, {x4-x12, x14} //stack stores the values of the arguments
@ -373,7 +374,6 @@ end_loop:
// d8 is used as dummy register and loaded along with d14 using ldp. d8 is not used in the function.
ldp d12,d13,[sp],#16
ldp d10,d11,[sp],#16
EXIT_FUNC
ret

6
common/arm64/ihevc_intra_pred_chroma_ver.s
View file

@ -87,15 +87,16 @@
// nt
// mode
.include "ihevc_neon_macros.s"
.text
.align 4
.include "ihevc_neon_macros.s"
.globl ihevc_intra_pred_chroma_ver_av8
.type ihevc_intra_pred_chroma_ver_av8, %function
ENTRY ihevc_intra_pred_chroma_ver_av8
ihevc_intra_pred_chroma_ver_av8:
// stmfd sp!, {x4-x12, x14} //stack stores the values of the arguments
push_v_regs
@ -225,7 +226,6 @@ end_func:
// ldmfd sp!,{x4-x12,x15} //reload the registers from sp
ldp x19, x20,[sp],#16
pop_v_regs
EXIT_FUNC
ret

6
common/arm64/ihevc_intra_pred_filters_chroma_mode_11_to_17.s
View file

@ -88,8 +88,9 @@
// nt
// mode
.include "ihevc_neon_macros.s"
.text
.align 4
.include "ihevc_neon_macros.s"
@ -101,7 +102,7 @@
.type ihevc_intra_pred_chroma_mode_11_to_17_av8, %function
ENTRY ihevc_intra_pred_chroma_mode_11_to_17_av8
ihevc_intra_pred_chroma_mode_11_to_17_av8:
// stmfd sp!, {x4-x12, x14} //stack stores the values of the arguments
@ -616,7 +617,6 @@ end_func:
ldp x19, x20,[sp],#16
ldp d14,d15,[sp],#16
ldp d12,d13,[sp],#16
EXIT_FUNC
ret

6
common/arm64/ihevc_intra_pred_filters_chroma_mode_19_to_25.s
View file

@ -88,8 +88,9 @@
// nt
// mode
.include "ihevc_neon_macros.s"
.text
.align 4
.include "ihevc_neon_macros.s"
.globl ihevc_intra_pred_chroma_mode_19_to_25_av8
@ -99,7 +100,7 @@
.type ihevc_intra_pred_chroma_mode_19_to_25_av8, %function
ENTRY ihevc_intra_pred_chroma_mode_19_to_25_av8
ihevc_intra_pred_chroma_mode_19_to_25_av8:
// stmfd sp!, {x4-x12, x14} //stack stores the values of the arguments
@ -570,7 +571,6 @@ end_loops:
ldp d8,d14,[sp],#16 // Loading d14 using { ldr d14,[sp]; add sp,sp,#8 } is giving bus error.
// d8 is used as dummy register and loaded along with d14 using ldp. d8 is not used in the function.
ldp d12,d13,[sp],#16
EXIT_FUNC
ret

6
common/arm64/ihevc_intra_pred_filters_luma_mode_11_to_17.s
View file

@ -88,8 +88,9 @@
// nt
// mode
.include "ihevc_neon_macros.s"
.text
.align 4
.include "ihevc_neon_macros.s"
@ -101,7 +102,7 @@
.type ihevc_intra_pred_luma_mode_11_to_17_av8, %function
ENTRY ihevc_intra_pred_luma_mode_11_to_17_av8
ihevc_intra_pred_luma_mode_11_to_17_av8:
// stmfd sp!, {x4-x12, x14} //stack stores the values of the arguments
@ -690,7 +691,6 @@ end_func:
ldp x19, x20,[sp],#16
ldp d14,d15,[sp],#16
ldp d12,d13,[sp],#16
EXIT_FUNC
ret

6
common/arm64/ihevc_intra_pred_filters_luma_mode_19_to_25.s
View file

@ -88,8 +88,9 @@
// nt
// mode
.include "ihevc_neon_macros.s"
.text
.align 4
.include "ihevc_neon_macros.s"
@ -100,7 +101,7 @@
.type ihevc_intra_pred_luma_mode_19_to_25_av8, %function
ENTRY ihevc_intra_pred_luma_mode_19_to_25_av8
ihevc_intra_pred_luma_mode_19_to_25_av8:
// stmfd sp!, {x4-x12, x14} //stack stores the values of the arguments
@ -660,7 +661,6 @@ end_loops:
ldp d14,d15,[sp],#16
ldp d12,d13,[sp],#16
ldp d9,d10,[sp],#16
EXIT_FUNC
ret

6
common/arm64/ihevc_intra_pred_luma_dc.s
View file

@ -92,15 +92,16 @@
// mode
// pi1_coeff
.include "ihevc_neon_macros.s"
.text
.align 4
.include "ihevc_neon_macros.s"
.globl ihevc_intra_pred_luma_dc_av8
.type ihevc_intra_pred_luma_dc_av8, %function
ENTRY ihevc_intra_pred_luma_dc_av8
ihevc_intra_pred_luma_dc_av8:
// stmfd sp!, {x4-x12, x14} //stack stores the values of the arguments
@ -510,7 +511,6 @@ end_func:
// ldmfd sp!,{x4-x12,x15} //reload the registers from sp
ldp x19, x20,[sp],#16
EXIT_FUNC
ret

9
common/arm64/ihevc_intra_pred_luma_horz.s
View file

@ -84,8 +84,9 @@
//x2 => *pu1_dst
//x3 => dst_strd
.include "ihevc_neon_macros.s"
.text
.align 4
.include "ihevc_neon_macros.s"
@ -93,7 +94,7 @@
.type ihevc_intra_pred_luma_horz_av8, %function
ENTRY ihevc_intra_pred_luma_horz_av8
ihevc_intra_pred_luma_horz_av8:
// stmfd sp!, {x4-x12, x14} //stack stores the values of the arguments
@ -188,7 +189,6 @@ core_loop_32:
// ldmfd sp!,{x4-x12,x15} //reload the registers from sp
ldp x19, x20,[sp],#16
EXIT_FUNC
ret
b end_func
@ -269,7 +269,6 @@ core_loop_16:
// ldmfd sp!,{x4-x12,x15} //reload the registers from sp
ldp x19, x20,[sp],#16
EXIT_FUNC
ret
b end_func
@ -316,7 +315,6 @@ core_loop_8:
// ldmfd sp!,{x4-x12,x15} //reload the registers from sp
ldp x19, x20,[sp],#16
EXIT_FUNC
ret
b end_func
@ -352,7 +350,6 @@ core_loop_4:
// ldmfd sp!,{x4-x12,x15} //reload the registers from sp
ldp x19, x20,[sp],#16
EXIT_FUNC
ret
end_func:

6
common/arm64/ihevc_intra_pred_luma_mode2.s
View file

@ -92,8 +92,9 @@
// mode
// pi1_coeff
.include "ihevc_neon_macros.s"
.text
.align 4
.include "ihevc_neon_macros.s"
@ -101,7 +102,7 @@
.type ihevc_intra_pred_luma_mode2_av8, %function
ENTRY ihevc_intra_pred_luma_mode2_av8
ihevc_intra_pred_luma_mode2_av8:
// stmfd sp!, {x4-x12, x14} //stack stores the values of the arguments
@ -269,7 +270,6 @@ end_func:
// ldmfd sp!,{x4-x12,x15} //reload the registers from sp
ldp x19, x20,[sp],#16
EXIT_FUNC
ret

6
common/arm64/ihevc_intra_pred_luma_mode_18_34.s
View file

@ -92,8 +92,9 @@
// mode
// pi1_coeff
.include "ihevc_neon_macros.s"
.text
.align 4
.include "ihevc_neon_macros.s"
@ -101,7 +102,7 @@
.type ihevc_intra_pred_luma_mode_18_34_av8, %function
ENTRY ihevc_intra_pred_luma_mode_18_34_av8
ihevc_intra_pred_luma_mode_18_34_av8:
// stmfd sp!, {x4-x12, x14} //stack stores the values of the arguments
push_v_regs
@ -277,7 +278,6 @@ end_func:
// ldmfd sp!,{x4-x12,x15} //reload the registers from sp
ldp x19, x20,[sp],#16
pop_v_regs
EXIT_FUNC
ret

6
common/arm64/ihevc_intra_pred_luma_mode_27_to_33.s
View file

@ -85,8 +85,9 @@
//x2 => *pu1_dst
//x3 => dst_strd
.include "ihevc_neon_macros.s"
.text
.align 4
.include "ihevc_neon_macros.s"
@ -96,7 +97,7 @@
.type ihevc_intra_pred_luma_mode_27_to_33_av8, %function
ENTRY ihevc_intra_pred_luma_mode_27_to_33_av8
ihevc_intra_pred_luma_mode_27_to_33_av8:
// stmfd sp!, {x4-x12, x14} //stack stores the values of the arguments
@ -553,7 +554,6 @@ end_loops:
ldp d14,d15,[sp],#16
ldp d12,d13,[sp],#16
ldp d9,d10,[sp],#16
EXIT_FUNC
ret

6
common/arm64/ihevc_intra_pred_luma_mode_3_to_9.s
View file

@ -88,8 +88,9 @@
// nt
// mode
.include "ihevc_neon_macros.s"
.text
.align 4
.include "ihevc_neon_macros.s"
@ -102,7 +103,7 @@
.type ihevc_intra_pred_luma_mode_3_to_9_av8, %function
ENTRY ihevc_intra_pred_luma_mode_3_to_9_av8
ihevc_intra_pred_luma_mode_3_to_9_av8:
// stmfd sp!, {x4-x12, x14} //stack stores the values of the arguments
@ -562,7 +563,6 @@ end_func:
ldp x19, x20,[sp],#16
ldp d14,d15,[sp],#16
ldp d12,d13,[sp],#16
EXIT_FUNC
ret

6
common/arm64/ihevc_intra_pred_luma_planar.s
View file

@ -92,8 +92,9 @@
// mode
// pi1_coeff
.include "ihevc_neon_macros.s"
.text
.align 4
.include "ihevc_neon_macros.s"
@ -103,7 +104,7 @@
.type ihevc_intra_pred_luma_planar_av8, %function
ENTRY ihevc_intra_pred_luma_planar_av8
ihevc_intra_pred_luma_planar_av8:
// stmfd sp!, {x4-x12, x14} //stack stores the values of the arguments
@ -557,7 +558,6 @@ end_loop:
// ldmfd sp!,{x4-x12,x15} //reload the registers from sp
ldp x19, x20,[sp],#16
EXIT_FUNC
ret

6
common/arm64/ihevc_intra_pred_luma_vert.s
View file

@ -88,8 +88,9 @@
// nt
// mode
.include "ihevc_neon_macros.s"
.text
.align 4
.include "ihevc_neon_macros.s"
@ -97,7 +98,7 @@
.type ihevc_intra_pred_luma_ver_av8, %function
ENTRY ihevc_intra_pred_luma_ver_av8
ihevc_intra_pred_luma_ver_av8:
// stmfd sp!, {x4-x12, x14} //stack stores the values of the arguments
@ -423,7 +424,6 @@ end_func:
// ldmfd sp!,{x4-x12,x15} //reload the registers from sp
ldp x19, x20,[sp],#16
EXIT_FUNC
ret

7
common/arm64/ihevc_itrans_recon_16x16.s
View file

@ -105,8 +105,10 @@
// x12
// x11
.include "ihevc_neon_macros.s"
.text
.align 4
.include "ihevc_neon_macros.s"
@ -121,7 +123,7 @@
.type ihevc_itrans_recon_16x16_av8, %function
ENTRY ihevc_itrans_recon_16x16_av8
ihevc_itrans_recon_16x16_av8:
ldr w11, [sp]
// stmfd sp!,{x4-x12,x14}
@ -1224,7 +1226,6 @@ skip_last8rows_stage2_kernel2:
// ldmfd sp!,{x4-x12,pc}
ldp x19, x20,[sp],#16
pop_v_regs
EXIT_FUNC
ret

6
common/arm64/ihevc_itrans_recon_32x32.s
View file

@ -124,8 +124,9 @@
//d5[2]= 43 d7[2]=9
//d5[3]= 38 d7[3]=4
.include "ihevc_neon_macros.s"
.text
.align 4
.include "ihevc_neon_macros.s"
@ -142,7 +143,7 @@
.type ihevc_itrans_recon_32x32_av8, %function
ENTRY ihevc_itrans_recon_32x32_av8
ihevc_itrans_recon_32x32_av8:
ldr w11, [sp]
@ -3041,7 +3042,6 @@ prediction_buffer:
// ldmfd sp!,{x0-x12,pc}
ldp x19, x20,[sp],#16
pop_v_regs
EXIT_FUNC
ret

7
common/arm64/ihevc_itrans_recon_4x4.s
View file

@ -100,8 +100,10 @@
// x6 => dst_strd
// x7 => zero_cols
.include "ihevc_neon_macros.s"
.text
.align 4
.include "ihevc_neon_macros.s"
.set shift_stage1_idct , 7
.set shift_stage2_idct , 12
@ -114,7 +116,7 @@
.type ihevc_itrans_recon_4x4_av8, %function
ENTRY ihevc_itrans_recon_4x4_av8
ihevc_itrans_recon_4x4_av8:
// stmfd sp!, {x4-x12, x14} //stack stores the values of the arguments
@ -227,7 +229,6 @@ ENTRY ihevc_itrans_recon_4x4_av8
// ldmfd sp!,{x4-x12,x15} //reload the registers from sp
ldp x19, x20,[sp],#16
EXIT_FUNC
ret

7
common/arm64/ihevc_itrans_recon_4x4_ttype1.s
View file

@ -103,8 +103,10 @@
// x6 => dst_strd
// x7 => zero_cols
.include "ihevc_neon_macros.s"
.text
.align 4
.include "ihevc_neon_macros.s"
.set shift_stage1_idct , 7
.set shift_stage2_idct , 12
@ -113,7 +115,7 @@
.type ihevc_itrans_recon_4x4_ttype1_av8, %function
ENTRY ihevc_itrans_recon_4x4_ttype1_av8
ihevc_itrans_recon_4x4_ttype1_av8:
// stmfd sp!, {x4-x12, x14} //stack stores the values of the arguments
@ -232,7 +234,6 @@ ENTRY ihevc_itrans_recon_4x4_ttype1_av8
// ldmfd sp!,{x4-x12,x15} //reload the registers from sp
ldp x19, x20,[sp],#16
EXIT_FUNC
ret

7
common/arm64/ihevc_itrans_recon_8x8.s
View file

@ -105,8 +105,10 @@
// zero_cols
.include "ihevc_neon_macros.s"
.text
.align 4
.include "ihevc_neon_macros.s"
@ -121,7 +123,7 @@
.type ihevc_itrans_recon_8x8_av8, %function
ENTRY ihevc_itrans_recon_8x8_av8
ihevc_itrans_recon_8x8_av8:
////register usage.extern - loading and until idct of columns
//// cosine constants - d0
//// sine constants - d1
@ -1028,7 +1030,6 @@ pred_buff_addition:
// ldmfd sp!,{x4-x12,pc}
ldp x19, x20,[sp],#16
pop_v_regs
EXIT_FUNC
ret

22
common/arm64/ihevc_mem_fns.s
View file

@ -69,14 +69,14 @@
// x1 => *pu1_src
// x2 => num_bytes
.include "ihevc_neon_macros.s"
.text
.p2align 2
.global ihevc_memcpy_mul_8_av8
.type ihevc_memcpy_mul_8_av8, %function
ENTRY ihevc_memcpy_mul_8_av8
ihevc_memcpy_mul_8_av8:
LOOP_NEON_MEMCPY_MUL_8:
// Memcpy 8 bytes
@ -85,7 +85,6 @@ LOOP_NEON_MEMCPY_MUL_8:
SUBS x2,x2,#8
BNE LOOP_NEON_MEMCPY_MUL_8
EXIT_FUNC
ret
@ -105,7 +104,7 @@ LOOP_NEON_MEMCPY_MUL_8:
.global ihevc_memcpy_av8
.type ihevc_memcpy_av8, %function
ENTRY ihevc_memcpy_av8
ihevc_memcpy_av8:
SUBS x2,x2,#8
BLT ARM_MEMCPY
LOOP_NEON_MEMCPY:
@ -127,7 +126,6 @@ LOOP_ARM_MEMCPY:
SUBS x2,x2,#1
BNE LOOP_ARM_MEMCPY
MEMCPY_RETURN:
EXIT_FUNC
ret
@ -142,13 +140,14 @@ MEMCPY_RETURN:
// x2 => num_bytes
.text
.p2align 2
.global ihevc_memset_mul_8_av8
.type ihevc_memset_mul_8_av8, %function
ENTRY ihevc_memset_mul_8_av8
ihevc_memset_mul_8_av8:
// Assumptions: numbytes is either 8, 16 or 32
dup v0.8b,w1
@ -159,7 +158,6 @@ LOOP_MEMSET_MUL_8:
SUBS x2,x2,#8
BNE LOOP_MEMSET_MUL_8
EXIT_FUNC
ret
@ -178,7 +176,7 @@ LOOP_MEMSET_MUL_8:
.global ihevc_memset_av8
.type ihevc_memset_av8, %function
ENTRY ihevc_memset_av8
ihevc_memset_av8:
SUBS x2,x2,#8
BLT ARM_MEMSET
dup v0.8b,w1
@ -200,7 +198,6 @@ LOOP_ARM_MEMSET:
BNE LOOP_ARM_MEMSET
MEMSET_RETURN:
EXIT_FUNC
ret
@ -215,13 +212,14 @@ MEMSET_RETURN:
// x2 => num_words
.text
.p2align 2
.global ihevc_memset_16bit_mul_8_av8
.type ihevc_memset_16bit_mul_8_av8, %function
ENTRY ihevc_memset_16bit_mul_8_av8
ihevc_memset_16bit_mul_8_av8:
// Assumptions: num_words is either 8, 16 or 32
@ -233,7 +231,6 @@ LOOP_MEMSET_16BIT_MUL_8:
SUBS x2,x2,#8
BNE LOOP_MEMSET_16BIT_MUL_8
EXIT_FUNC
ret
@ -252,7 +249,7 @@ LOOP_MEMSET_16BIT_MUL_8:
.global ihevc_memset_16bit_av8
.type ihevc_memset_16bit_av8, %function
ENTRY ihevc_memset_16bit_av8
ihevc_memset_16bit_av8:
SUBS x2,x2,#8
BLT ARM_MEMSET_16BIT
dup v0.8h,w1
@ -274,7 +271,6 @@ LOOP_ARM_MEMSET_16BIT:
BNE LOOP_ARM_MEMSET_16BIT
MEMSET_16BIT_RETURN:
EXIT_FUNC
ret

50
common/arm64/ihevc_neon_macros.s
View file

@ -47,53 +47,3 @@
ldp d10,d11,[sp],#16
ldp d8,d9,[sp],#16
.endm
// --- Internal Security Dispatchers ---
// These expand to real instructions only if the compiler flags are present.
.macro BTI_ENABLE
#if defined(__ARM_FEATURE_BTI_DEFAULT)
bti c
#endif
.endm
.macro PAC_ENTRY
#if defined(__ARM_FEATURE_PAC_DEFAULT)
paciasp
#endif
.endm
.macro PAC_EXIT
#if defined(__ARM_FEATURE_PAC_DEFAULT)
autiasp
#endif
.endm
// --- Main ENTRY and EXIT_FUNC Macros ---
.macro ENTRY name
.p2align 2
\name:
BTI_ENABLE
PAC_ENTRY
.endm
.macro EXIT_FUNC
PAC_EXIT
.endm
// --- GNU Property Note ---
// Signals BTI and PAC support to the Android linker.
#if defined(__linux__) && defined(__aarch64__)
.pushsection .note.gnu.property, "a" // Switch to Note section
.p2align 3
.word 4 // Name size
.word 16 // Data size
.word 5 // NT_GNU_PROPERTY_TYPE_0
.asciz "GNU" // Owner
.word 0xc0000000 // GNU_PROPERTY_AARCH64_FEATURE_1_AND
.word 4 // Data size
.word 3 // Value: BTI (Bit 0) | PAC (Bit 1)
.word 0 // Padding
.popsection // Switch back to previous section
#endif

14
common/arm64/ihevc_padding.s
View file

@ -85,14 +85,14 @@
// x2 => ht
// x3 => pad_size
.include "ihevc_neon_macros.s"
.text
.align 4
.globl ihevc_pad_left_luma_av8
.type ihevc_pad_left_luma_av8, %function
ENTRY ihevc_pad_left_luma_av8
ihevc_pad_left_luma_av8:
loop_start_luma_left:
// pad size is assumed to be pad_left = 80
@ -148,7 +148,6 @@ loop_start_luma_left:
bne loop_start_luma_left
EXIT_FUNC
ret
@ -210,7 +209,7 @@ loop_start_luma_left:
.type ihevc_pad_left_chroma_av8, %function
ENTRY ihevc_pad_left_chroma_av8
ihevc_pad_left_chroma_av8:
loop_start_chroma_left:
@ -267,7 +266,6 @@ loop_start_chroma_left:
bne loop_start_chroma_left
EXIT_FUNC
ret
@ -339,7 +337,7 @@ loop_start_chroma_left:
.type ihevc_pad_right_luma_av8, %function
ENTRY ihevc_pad_right_luma_av8
ihevc_pad_right_luma_av8:
loop_start_luma_right:
@ -397,7 +395,6 @@ loop_start_luma_right:
bne loop_start_luma_right
EXIT_FUNC
ret
@ -458,7 +455,7 @@ loop_start_luma_right:
.type ihevc_pad_right_chroma_av8, %function
ENTRY ihevc_pad_right_chroma_av8
ihevc_pad_right_chroma_av8:
loop_start_chroma_right:
@ -515,7 +512,6 @@ loop_start_chroma_right:
bne loop_start_chroma_right
EXIT_FUNC
ret

6
common/arm64/ihevc_sao_band_offset_chroma.s
View file

@ -61,13 +61,14 @@
//x9 => wd 60
//x10=> ht 64
.include "ihevc_neon_macros.s"
.text
.p2align 2
.include "ihevc_neon_macros.s"
.globl gu1_table_band_idx
.globl ihevc_sao_band_offset_chroma_av8
ENTRY ihevc_sao_band_offset_chroma_av8
ihevc_sao_band_offset_chroma_av8:
mov x8,#0
mov x9,#0
mov x10,#0
@ -423,7 +424,6 @@ END_LOOP:
ldp x21, x22,[sp],#16
ldp x19, x20,[sp],#16
pop_v_regs
EXIT_FUNC
ret

4
common/arm64/ihevc_sao_band_offset_luma.s
View file

@ -65,11 +65,12 @@
.include "ihevc_neon_macros.s"
.text
.p2align 2
.globl gu1_table_band_idx
.globl ihevc_sao_band_offset_luma_av8
ENTRY ihevc_sao_band_offset_luma_av8
ihevc_sao_band_offset_luma_av8:
// STMFD sp!, {x4-x12, x14} //stack stores the values of the arguments
@ -243,7 +244,6 @@ HEIGHT_LOOP:
ldp d8,d15,[sp],#16 // Loading d15 using { ldr d15,[sp]; add sp,sp,#8 } is giving bus error.
// d8 is used as dummy register and loaded along with d15 using ldp. d8 is not used in the function.
ldp d13,d14,[sp],#16
EXIT_FUNC
ret

6
common/arm64/ihevc_sao_edge_offset_class0.s
View file

@ -59,14 +59,15 @@
//x9 => wd
//x10=> ht
.text
.p2align 2
.include "ihevc_neon_macros.s"
.text
.globl gi1_table_edge_idx
.globl ihevc_sao_edge_offset_class0_av8
ENTRY ihevc_sao_edge_offset_class0_av8
ihevc_sao_edge_offset_class0_av8:
// STMFD sp!, {x4-x12, x14} //stack stores the values of the arguments
@ -337,7 +338,6 @@ END_LOOPS:
// LDMFD sp!,{x4-x12,x15} //Reload the registers from SP
ldp x19, x20,[sp], #16
EXIT_FUNC
ret

6
common/arm64/ihevc_sao_edge_offset_class0_chroma.s
View file

@ -60,13 +60,14 @@
//x9 => wd
//x10=> ht
.include "ihevc_neon_macros.s"
.text
.p2align 2
.include "ihevc_neon_macros.s"
.globl gi1_table_edge_idx
.globl ihevc_sao_edge_offset_class0_chroma_av8
ENTRY ihevc_sao_edge_offset_class0_chroma_av8
ihevc_sao_edge_offset_class0_chroma_av8:
ldr x8,[sp,#0]
ldr x9,[sp,#8]
@ -476,7 +477,6 @@ END_LOOPS:
ldp x21, x22,[sp],#16
ldp x19, x20,[sp],#16
EXIT_FUNC
ret

6
common/arm64/ihevc_sao_edge_offset_class1.s
View file

@ -58,14 +58,15 @@
//x7 => wd
//x8 => ht
.text
.p2align 2
.include "ihevc_neon_macros.s"
.text
.globl gi1_table_edge_idx
.globl ihevc_sao_edge_offset_class1_av8
ENTRY ihevc_sao_edge_offset_class1_av8
ihevc_sao_edge_offset_class1_av8:
// STMFD sp!, {x4-x12, x14} //stack stores the values of the arguments
@ -354,7 +355,6 @@ END_LOOPS:
// LDMFD sp!,{x4-x12,x15} //Reload the registers from SP
ldp x19, x20,[sp], #16
EXIT_FUNC
ret

6
common/arm64/ihevc_sao_edge_offset_class1_chroma.s
View file

@ -60,13 +60,14 @@
//x8 => wd
//x9 => ht
.include "ihevc_neon_macros.s"
.text
.p2align 2
.include "ihevc_neon_macros.s"
.globl gi1_table_edge_idx
.globl ihevc_sao_edge_offset_class1_chroma_av8
ENTRY ihevc_sao_edge_offset_class1_chroma_av8
ihevc_sao_edge_offset_class1_chroma_av8:
ldr x8,[sp,#0]
@ -457,7 +458,6 @@ END_LOOPS:
ldp x19, x20,[sp],#16
EXIT_FUNC
ret

6
common/arm64/ihevc_sao_edge_offset_class2.s
View file

@ -58,14 +58,15 @@
//x7 => wd
//x8=> ht
.text
.p2align 2
.include "ihevc_neon_macros.s"
.text
.globl gi1_table_edge_idx
.globl ihevc_sao_edge_offset_class2_av8
ENTRY ihevc_sao_edge_offset_class2_av8
ihevc_sao_edge_offset_class2_av8:
// STMFD sp!,{x4-x12,x14} //stack stores the values of the arguments
@ -845,7 +846,6 @@ END_LOOPS:
ldp x21, x22,[sp],#16
ldp x19, x20,[sp],#16
EXIT_FUNC
ret

6
common/arm64/ihevc_sao_edge_offset_class2_chroma.s
View file

@ -60,13 +60,14 @@
//x7 => wd
//x8=> ht
.include "ihevc_neon_macros.s"
.text
.p2align 2
.include "ihevc_neon_macros.s"
.globl gi1_table_edge_idx
.globl ihevc_sao_edge_offset_class2_chroma_av8
ENTRY ihevc_sao_edge_offset_class2_chroma_av8
ihevc_sao_edge_offset_class2_chroma_av8:
// STMFD sp!,{x4-x12,x14} //stack stores the values of the arguments
@ -1131,7 +1132,6 @@ END_LOOPS:
ldp x21, x22,[sp],#16
ldp x19, x20,[sp],#16
EXIT_FUNC
ret

6
common/arm64/ihevc_sao_edge_offset_class3.s
View file

@ -58,14 +58,15 @@
//x7 => wd
//x8=> ht
.text
.p2align 2
.include "ihevc_neon_macros.s"
.text
.globl gi1_table_edge_idx
.globl ihevc_sao_edge_offset_class3_av8
ENTRY ihevc_sao_edge_offset_class3_av8
ihevc_sao_edge_offset_class3_av8:
// STMFD sp!,{x4-x12,x14} //stack stores the values of the arguments
@ -888,7 +889,6 @@ END_LOOPS:
ldp x23, x24,[sp], #16
ldp x21, x22,[sp], #16
ldp x19, x20,[sp], #16
EXIT_FUNC
ret

6
common/arm64/ihevc_sao_edge_offset_class3_chroma.s
View file

@ -60,12 +60,13 @@
//x7 => wd
//x8=> ht
.include "ihevc_neon_macros.s"
.text
.p2align 2
.include "ihevc_neon_macros.s"
.globl gi1_table_edge_idx
.globl ihevc_sao_edge_offset_class3_chroma_av8
ENTRY ihevc_sao_edge_offset_class3_chroma_av8
ihevc_sao_edge_offset_class3_chroma_av8:
// STMFD sp!,{x4-x12,x14} //stack stores the values of the arguments
@ -1166,7 +1167,6 @@ END_LOOPS:
ldp x21, x22,[sp],#16
ldp x19, x20,[sp],#16
EXIT_FUNC
ret

7
common/arm64/ihevc_weighted_pred_bi.s
View file

@ -134,14 +134,16 @@
// x14 => ht
// x7 => wd
.include "ihevc_neon_macros.s"
.text
.align 4
.include "ihevc_neon_macros.s"
.globl ihevc_weighted_pred_bi_av8
.type ihevc_weighted_pred_bi_av8, %function
ENTRY ihevc_weighted_pred_bi_av8
ihevc_weighted_pred_bi_av8:
// stmfd sp!, {x4-x12, x14} //stack stores the values of the arguments
@ -305,7 +307,6 @@ end_loops:
ldp x21, x22,[sp],#16
ldp x19, x20,[sp],#16
EXIT_FUNC
ret

6
common/arm64/ihevc_weighted_pred_bi_default.s
View file

@ -107,15 +107,16 @@
// x7 => lvl_shift2
// x8 => ht
// x9 => wd
.text
.align 4
.include "ihevc_neon_macros.s"
.text
.globl ihevc_weighted_pred_bi_default_av8
.type ihevc_weighted_pred_bi_default_av8, %function
ENTRY ihevc_weighted_pred_bi_default_av8
ihevc_weighted_pred_bi_default_av8:
ldr w8,[sp,#0]
ldr w9,[sp,#8]
@ -533,7 +534,6 @@ end_loops:
ldp x21, x22,[sp],#16
ldp x19, x20,[sp],#16
EXIT_FUNC
ret

7
common/arm64/ihevc_weighted_pred_uni.s
View file

@ -112,14 +112,16 @@
// x8 => ht
// x9 => wd
.include "ihevc_neon_macros.s"
.text
.align 4
.include "ihevc_neon_macros.s"
.globl ihevc_weighted_pred_uni_av8
.type ihevc_weighted_pred_uni_av8, %function
ENTRY ihevc_weighted_pred_uni_av8
ihevc_weighted_pred_uni_av8:
// stmfd sp!, {x4-x12, x14} //stack stores the values of the arguments
@ -238,7 +240,6 @@ end_loops:
ldp x21, x22,[sp],#16
ldp x19, x20,[sp],#16
EXIT_FUNC
ret

6
common/common.cmake
View file

@ -18,7 +18,6 @@ list(
"${HEVC_ROOT}/common/ihevc_trans_tables.c"
"${HEVC_ROOT}/common/ihevc_recon.c"
"${HEVC_ROOT}/common/ihevc_itrans.c"
"${HEVC_ROOT}/common/ihevc_itrans_res.c"
"${HEVC_ROOT}/common/ihevc_itrans_recon.c"
"${HEVC_ROOT}/common/ihevc_iquant_recon.c"
"${HEVC_ROOT}/common/ihevc_iquant_itrans_recon.c"
@ -30,7 +29,6 @@ list(
"${HEVC_ROOT}/common/ihevc_chroma_iquant_itrans_recon.c"
"${HEVC_ROOT}/common/ihevc_chroma_recon.c"
"${HEVC_ROOT}/common/ihevc_chroma_itrans_recon_16x16.c"
"${HEVC_ROOT}/common/ihevc_chroma_itrans_recon_32x32.c"
"${HEVC_ROOT}/common/ihevc_chroma_itrans_recon_8x8.c"
"${HEVC_ROOT}/common/ihevc_buf_mgr.c"
"${HEVC_ROOT}/common/ihevc_disp_mgr.c"
@ -64,7 +62,7 @@ list(
include_directories(${HEVC_ROOT}/common)
# arm/x86 sources
if("${SYSTEM_PROCESSOR}" STREQUAL "aarch64" OR "${SYSTEM_PROCESSOR}" STREQUAL "arm64")
if("${CMAKE_SYSTEM_PROCESSOR}" STREQUAL "aarch64")
list(
APPEND
LIBHEVC_COMMON_ASMS
@ -137,7 +135,7 @@ if("${SYSTEM_PROCESSOR}" STREQUAL "aarch64" OR "${SYSTEM_PROCESSOR}" STREQUAL "a
"${HEVC_ROOT}/common/arm64/ihevc_weighted_pred_uni.s")
include_directories(${HEVC_ROOT}/common/arm64 ${HEVC_ROOT}/common/arm)
elseif("${SYSTEM_PROCESSOR}" STREQUAL "aarch32")
elseif("${CMAKE_SYSTEM_PROCESSOR}" STREQUAL "aarch32")
list(
APPEND
LIBHEVC_COMMON_ASMS

1562
common/ihevc_cabac_tables.c
View file

File diff suppressed because it is too large Load diff

8
common/ihevc_cabac_tables.h
View file

@ -92,13 +92,9 @@ typedef enum
IHEVC_CAB_COEFFY_PREFIX = IHEVC_CAB_COEFFX_PREFIX + 18,
IHEVC_CAB_CODED_SUBLK_IDX = IHEVC_CAB_COEFFY_PREFIX + 18,
IHEVC_CAB_COEFF_FLAG = IHEVC_CAB_CODED_SUBLK_IDX + 4,
IHEVC_CAB_COEFABS_GRTR1_FLAG = IHEVC_CAB_COEFF_FLAG + 44,
IHEVC_CAB_COEFABS_GRTR1_FLAG = IHEVC_CAB_COEFF_FLAG + 42,
IHEVC_CAB_COEFABS_GRTR2_FLAG = IHEVC_CAB_COEFABS_GRTR1_FLAG + 24,
IHEVC_CAB_CCP_LOG2_RES_ABS = IHEVC_CAB_COEFABS_GRTR2_FLAG + 6,
IHEVC_CAB_CCP_RES_SIGN_FLAG = IHEVC_CAB_CCP_LOG2_RES_ABS + 8,
IHEVC_CAB_EXPLICIT_RDPCM_FLAG = IHEVC_CAB_CCP_RES_SIGN_FLAG + 2,
IHEVC_CAB_EXPLICIT_RDPCM_DIR = IHEVC_CAB_EXPLICIT_RDPCM_FLAG + 2,
IHEVC_CAB_CTXT_END = IHEVC_CAB_EXPLICIT_RDPCM_DIR + 2,
IHEVC_CAB_CTXT_END = IHEVC_CAB_COEFABS_GRTR2_FLAG + 6
}IHEVC_CABAC_CTXT_OFFSETS;

10
common/ihevc_chroma_intra_pred.h
View file

@ -126,14 +126,7 @@ typedef void ihevc_intra_pred_chroma_ref_substitution_ft(UWORD8 *pu1_top_left,
WORD32 nt,
WORD32 nbr_flags,
UWORD8 *pu1_dst,
WORD32 dst_strd,
WORD32 chroma_format_idc);
typedef void ihevc_intra_pred_chroma_ref_filtering_ft(UWORD8 *pu1_src,
WORD32 nt,
UWORD8 *pu1_dst,
WORD32 mode,
WORD32 intra_smoothing_flags);
WORD32 dst_strd);
typedef void ihevc_hbd_intra_pred_chroma_planar_ft(
UWORD16 *pu2_ref,
@ -247,7 +240,6 @@ ihevc_intra_pred_chroma_mode_11_to_17_ft ihevc_intra_pred_chroma_mode_11_to_17;
ihevc_intra_pred_chroma_mode_19_to_25_ft ihevc_intra_pred_chroma_mode_19_to_25;
ihevc_intra_pred_chroma_mode_27_to_33_ft ihevc_intra_pred_chroma_mode_27_to_33;
ihevc_intra_pred_chroma_ref_substitution_ft ihevc_intra_pred_chroma_ref_substitution;
ihevc_intra_pred_chroma_ref_filtering_ft ihevc_intra_pred_chroma_ref_filtering;
ihevc_hbd_intra_pred_chroma_planar_ft ihevc_hbd_intra_pred_chroma_planar;
ihevc_hbd_intra_pred_chroma_dc_ft ihevc_hbd_intra_pred_chroma_dc;

141
common/ihevc_chroma_intra_pred_filters.c
View file

@ -64,7 +64,6 @@
#include "ihevc_typedefs.h"
#include "ihevc_macros.h"
#include "ihevc_defs.h"
#include "ihevc_func_selector.h"
#include "ihevc_platform_macros.h"
#include "ihevc_intra_pred.h"
@ -130,8 +129,7 @@ void ihevc_intra_pred_chroma_ref_substitution(UWORD8 *pu1_top_left,
WORD32 nt,
WORD32 nbr_flags,
UWORD8 *pu1_dst,
WORD32 dst_strd,
WORD32 chroma_format_idc)
WORD32 dst_strd)
{
UWORD8 pu1_ref_u, pu1_ref_v;
WORD32 dc_val, i, j;
@ -208,7 +206,7 @@ void ihevc_intra_pred_chroma_ref_substitution(UWORD8 *pu1_top_left,
// U-V interleaved Top-top right samples
}
if(nt == 4 || (nt == 8 && chroma_format_idc == CHROMA_FMT_IDC_YUV444))
if(nt == 4)
{
/* 1 bit extraction for all the neighboring blocks */
tp_left = (nbr_flags & 0x10000) >> 16;
@ -276,9 +274,8 @@ void ihevc_intra_pred_chroma_ref_substitution(UWORD8 *pu1_top_left,
}
}
else if(nt == 8 || (nt == 16 && chroma_format_idc == CHROMA_FMT_IDC_YUV444))
else if(nt == 8)
{
WORD32 sub_sample = chroma_format_idc == CHROMA_FMT_IDC_YUV444 ? 2 : 1;
WORD32 nbr_flags_temp = 0;
nbr_flags_temp = ((nbr_flags & 0xC) >> 2) + ((nbr_flags & 0xC0) >> 4)
+ ((nbr_flags & 0x300) >> 4)
@ -288,16 +285,16 @@ void ihevc_intra_pred_chroma_ref_substitution(UWORD8 *pu1_top_left,
/* compute trailing zeors based on nbr_flag for substitution process of below left see section .*/
/* as each bit in nbr flags corresponds to 8 pels for bot_left, left, top and topright but 1 pel for topleft */
{
nbr_id_from_bl = look_up_trailing_zeros(nbr_flags_temp & 0XF) * (4 * sub_sample); /* for bottom left and left */
if(nbr_id_from_bl == 32 * sub_sample)
nbr_id_from_bl = 16 * sub_sample;
if(nbr_id_from_bl == 16 * sub_sample)
nbr_id_from_bl = look_up_trailing_zeros(nbr_flags_temp & 0XF) * 4; /* for bottom left and left */
if(nbr_id_from_bl == 32)
nbr_id_from_bl = 16;
if(nbr_id_from_bl == 16)
{
/* for top left : 1 pel per nbr bit */
if(!((nbr_flags_temp >> 8) & 0x1))
{
nbr_id_from_bl++;
nbr_id_from_bl += look_up_trailing_zeros((nbr_flags_temp >> 4) & 0xF) * 4 * sub_sample; /* top and top right; 8 pels per nbr bit */
nbr_id_from_bl += look_up_trailing_zeros((nbr_flags_temp >> 4) & 0xF) * 4; /* top and top right; 8 pels per nbr bit */
}
}
@ -316,14 +313,14 @@ void ihevc_intra_pred_chroma_ref_substitution(UWORD8 *pu1_top_left,
}
/* for the loop of 4*Nt+1 pixels (excluding pixels computed from reverse substitution) */
while(nbr_id_from_bl < ((T8C_4NT * sub_sample)+1))
while(nbr_id_from_bl < ((T8C_4NT)+1))
{
/* To Obtain the next unavailable idx flag after reverse neighbor substitution */
/* Divide by 8 to obtain the original index */
frwd_nbr_flag = (nbr_id_from_bl >> (chroma_format_idc == CHROMA_FMT_IDC_YUV444 ? 3 : 2)); /*+ (nbr_id_from_bl & 0x1);*/
frwd_nbr_flag = (nbr_id_from_bl >> 2); /*+ (nbr_id_from_bl & 0x1);*/
/* The Top-left flag is at the last bit location of nbr_flags*/
if(nbr_id_from_bl == (T8C_4NT * sub_sample / 2))
if(nbr_id_from_bl == (T8C_4NT / 2))
{
get_bits = GET_BIT(nbr_flags_temp, 8);
@ -342,23 +339,22 @@ void ihevc_intra_pred_chroma_ref_substitution(UWORD8 *pu1_top_left,
UWORD16 *pu2_dst;
/* 8 pel substitution (other than TL) */
pu2_dst = (UWORD16 *)&pu1_dst[(2 * nbr_id_from_bl) - 2];
ihevc_memset_16bit((UWORD16 *)(pu1_dst + (2 * nbr_id_from_bl)), pu2_dst[0], 4 * sub_sample);
ihevc_memset_16bit((UWORD16 *)(pu1_dst + (2 * nbr_id_from_bl)), pu2_dst[0], 4);
}
}
nbr_id_from_bl += (nbr_id_from_bl == (T8C_4NT * sub_sample / 2)) ? 1 : 4 * sub_sample;
nbr_id_from_bl += (nbr_id_from_bl == (T8C_4NT / 2)) ? 1 : 4;
}
}
else if(nt == 16 || (nt == 32 && chroma_format_idc == CHROMA_FMT_IDC_YUV444))
else if(nt == 16)
{
WORD32 sub_sample = chroma_format_idc == CHROMA_FMT_IDC_YUV444 ? 2 : 1;
/* compute trailing ones based on mbr_flag for substitution process of below left see section .*/
/* as each bit in nbr flags corresponds to 4 pels for bot_left, left, top and topright but 1 pel for topleft */
{
nbr_id_from_bl = look_up_trailing_zeros((nbr_flags & 0XFF)) * 4 * sub_sample; /* for bottom left and left */
nbr_id_from_bl = look_up_trailing_zeros((nbr_flags & 0XFF)) * 4; /* for bottom left and left */
if(nbr_id_from_bl == 32 * sub_sample)
if(nbr_id_from_bl == 32)
{
/* for top left : 1 pel per nbr bit */
if(!((nbr_flags >> 16) & 0x1))
@ -366,7 +362,7 @@ void ihevc_intra_pred_chroma_ref_substitution(UWORD8 *pu1_top_left,
/* top left not available */
nbr_id_from_bl++;
/* top and top right; 4 pels per nbr bit */
nbr_id_from_bl += look_up_trailing_zeros((nbr_flags >> 8) & 0xFF) * 4 * sub_sample;
nbr_id_from_bl += look_up_trailing_zeros((nbr_flags >> 8) & 0xFF) * 4;
}
}
/* Reverse Substitution Process*/
@ -384,14 +380,14 @@ void ihevc_intra_pred_chroma_ref_substitution(UWORD8 *pu1_top_left,
}
/* for the loop of 4*Nt+1 pixels (excluding pixels computed from reverse substitution) */
while(nbr_id_from_bl < ((T16C_4NT * sub_sample)+1))
while(nbr_id_from_bl < ((T16C_4NT)+1))
{
/* To Obtain the next unavailable idx flag after reverse neighbor substitution */
/* Devide by 4 to obtain the original index */
frwd_nbr_flag = (nbr_id_from_bl >> (chroma_format_idc == CHROMA_FMT_IDC_YUV444 ? 3 : 2)); /*+ (nbr_id_from_bl & 0x1);*/
frwd_nbr_flag = (nbr_id_from_bl >> 2); /*+ (nbr_id_from_bl & 0x1);*/
/* The Top-left flag is at the last bit location of nbr_flags*/
if(nbr_id_from_bl == (T16C_4NT * sub_sample / 2))
if(nbr_id_from_bl == (T16C_4NT / 2))
{
get_bits = GET_BIT(nbr_flags, 16);
/* only pel substitution for TL */
@ -409,110 +405,17 @@ void ihevc_intra_pred_chroma_ref_substitution(UWORD8 *pu1_top_left,
UWORD16 *pu2_dst;
/* 4 pel substitution (other than TL) */
pu2_dst = (UWORD16 *)&pu1_dst[(2 * nbr_id_from_bl) - 2];
ihevc_memset_16bit((UWORD16 *)(pu1_dst + (2 * nbr_id_from_bl)), pu2_dst[0], 4 * sub_sample);
ihevc_memset_16bit((UWORD16 *)(pu1_dst + (2 * nbr_id_from_bl)), pu2_dst[0], 4);
}
}
nbr_id_from_bl += (nbr_id_from_bl == (T16C_4NT * sub_sample / 2)) ? 1 : 4 * sub_sample;
nbr_id_from_bl += (nbr_id_from_bl == (T16C_4NT / 2)) ? 1 : 4;
}
}
}
}
/**
*******************************************************************************
*
* @brief
* Intra prediction interpolation filter for chroma ref_filtering (4:4:4)
*
*
* @par Description:
* Reference DC filtering for neighboring chroma samples dependent on TU size and
* mode Refer to section 8.4.4.2.3 in the standard
*
* @param[in] pu1_src
* UWORD8 pointer to the source
*
* @param[out] pu1_dst
* UWORD8 pointer to the destination
*
* @param[in] nt
* integer Transform Block size
*
* @param[in] mode
* integer intraprediction mode
*
* @param[in] strong_intra_smoothing_enable_flag
* integer containing intra_smoothing_disabled_flag and strong_smoothing_enable_flag
*
* @returns
*
* @remarks
* None
*
*******************************************************************************
*/
void ihevc_intra_pred_chroma_ref_filtering(UWORD8 *pu1_src,
WORD32 nt,
UWORD8 *pu1_dst,
WORD32 mode,
WORD32 intra_smoothing_flag)
{
WORD32 filter_flag;
WORD32 i; /* Generic indexing variable */
WORD32 four_nt = 4 * nt;
UWORD8 au1_flt[((4 * MAX_CU_SIZE) + 1) * 2];
WORD32 intra_smoothing_disabled_flag = (intra_smoothing_flag >> 3) & 0x1;
WORD32 strong_intra_smoothing_enable_flag = intra_smoothing_flag & 0x1;
UNUSED(strong_intra_smoothing_enable_flag);
if(intra_smoothing_disabled_flag)
{
if(pu1_src == pu1_dst) return;
for(i = 0; i < (2 * (four_nt + 1)); i += 2)
{
pu1_dst[i] = pu1_src[i];
pu1_dst[i + 1] = pu1_src[i + 1];
}
return;
}
filter_flag = gau1_intra_pred_ref_filter[mode] & (1 << (CTZ(nt) - 2));
if(0 == filter_flag)
{
if(pu1_src == pu1_dst) return;
for(i = 0; i < (2 * (four_nt + 1)); i += 2)
{
pu1_dst[i] = pu1_src[i];
pu1_dst[i + 1] = pu1_src[i + 1];
}
}
else
{
/* Extremities Untouched*/
au1_flt[0] = pu1_src[0];
au1_flt[1] = pu1_src[1];
au1_flt[four_nt * 2] = pu1_src[four_nt * 2];
au1_flt[(four_nt * 2) + 1] = pu1_src[(four_nt * 2) + 1];
for(i = 2; i < four_nt * 2; i += 2)
{
au1_flt[i] = (pu1_src[i - 2] + 2 * pu1_src[i] + pu1_src[i + 2] + 2) >> 2;
au1_flt[i + 1] = (pu1_src[i - 1] + 2 * pu1_src[i + 1] + pu1_src[i + 3] + 2) >> 2;
}
for(i = 0; i < (2 * (four_nt + 1)); i += 2)
{
pu1_dst[i] = au1_flt[i];
pu1_dst[i + 1] = au1_flt[i + 1];
}
}
}
/**
*******************************************************************************
*

10
common/ihevc_chroma_itrans_recon.h
View file

@ -94,15 +94,6 @@ typedef void ihevc_hbd_chroma_itrans_recon_16x16_ft(WORD16 *pi2_src,
WORD32 zero_cols,
WORD32 zero_rows,
UWORD8 bit_depth);
typedef void ihevc_chroma_itrans_recon_32x32_ft(WORD16 *pi2_src,
WORD16 *pi2_tmp,
UWORD8 *pu1_pred,
UWORD8 *pu1_dst,
WORD32 src_strd,
WORD32 pred_strd,
WORD32 dst_strd,
WORD32 zero_cols,
WORD32 zero_rows);
ihevc_chroma_itrans_recon_4x4_ft ihevc_chroma_itrans_recon_4x4;
ihevc_hbd_chroma_itrans_recon_4x4_ft ihevc_hbd_chroma_itrans_recon_4x4;
@ -110,7 +101,6 @@ ihevc_chroma_itrans_recon_8x8_ft ihevc_chroma_itrans_recon_8x8;
ihevc_hbd_chroma_itrans_recon_8x8_ft ihevc_hbd_chroma_itrans_recon_8x8;
ihevc_chroma_itrans_recon_16x16_ft ihevc_chroma_itrans_recon_16x16;
ihevc_hbd_chroma_itrans_recon_16x16_ft ihevc_hbd_chroma_itrans_recon_16x16;
ihevc_chroma_itrans_recon_32x32_ft ihevc_chroma_itrans_recon_32x32;
/* A9 Q Function Declarations */
ihevc_chroma_itrans_recon_4x4_ft ihevc_chroma_itrans_recon_4x4_a9q;

1135
common/ihevc_chroma_itrans_recon_32x32.c
View file

File diff suppressed because it is too large Load diff

154
common/ihevc_chroma_recon.c
View file

@ -306,157 +306,3 @@ void ihevc_chroma_recon_16x16(WORD16 *pi2_src,
}
}
/**
*******************************************************************************
*
* @brief
* This function performs reconstruction for 32x32 input block
*
* @par Description:
* Performs reconstruction of 32x32 input block by adding adding prediction
* data to input and clipping it to 8 bit
*
* @param[in] pi2_src
* Input 32x32 coefficients
*
* @param[in] pu1_pred
* Prediction 32x32 block
*
* @param[out] pu1_dst
* Output 32x32 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_tmp
*
* @returns Void
*
* @remarks
* None
*
*******************************************************************************
*/
void ihevc_chroma_recon_32x32(WORD16 *pi2_src,
UWORD8 *pu1_pred,
UWORD8 *pu1_dst,
WORD32 src_strd,
WORD32 pred_strd,
WORD32 dst_strd,
WORD32 zero_cols)
{
WORD32 i, j;
WORD32 trans_size;
trans_size = TRANS_SIZE_32;
/* Reconstruction */
for(i = 0; i < trans_size; i++)
{
/* Checking for Zero Cols */
if((zero_cols & 1) == 1)
{
for(j = 0; j < trans_size; j++)
{
pu1_dst[j * dst_strd] = pu1_pred[j * pred_strd];
}
}
else
{
for(j = 0; j < trans_size; j++)
{
pu1_dst[j * dst_strd] =
CLIP_U8(pi2_src[j * src_strd] + pu1_pred[j * pred_strd]);
}
}
pi2_src++;
pu1_dst += 2;
pu1_pred += 2;
zero_cols = zero_cols >> 1;
}
}
/**
******************************************************************************
*
* @brief Constructs chroma recon with Cross Component Prediction (CCP)
*
* @par Description
* This routine uses reconstructed luma residual samples to predict chroma
* residual samples as per HEVC Specification Section 8.6.6. It scales the
* luma residual by a signaled alpha value and adds it to the chroma residual
* prior to final reconstruction.
*
* @param[in] pi2_luma_res
* pointer to the luma residual
*
* @param[in] pi2_chroma_res
* pointer to the chroma residual
*
* @param[in] pu1_pred
* prediction block
*
* @param[in] pu1_dst
* destination block
*
* @param[in] alpha
* scaling factor for the luma residual
*
* @param[in] trans_size
* transform size
*
* @param[in] luma_res_stride
* stride of the luma residual buffer
*
* @param[in] chroma_res_stride
* stride of the chroma residual buffer
*
* @param[in] pred_strd
* Prediction stride
*
* @param[in] dst_strd
* Output Stride
*
* @return success or failure error code
*
******************************************************************************
*/
void ihevc_chroma_recon_nxn_ccp(WORD16 *pi2_luma_res,
WORD16 *pi2_chroma_res,
UWORD8 *pu1_pred,
UWORD8 *pu1_dst,
WORD32 alpha,
WORD32 trans_size,
WORD32 luma_res_stride,
WORD32 chroma_res_stride,
WORD32 pred_stride,
WORD32 dst_stride)
{
WORD32 i, j;
for(i = 0; i < trans_size; i++)
{
for(j = 0; j < trans_size; j++)
{
WORD32 res = (alpha * pi2_luma_res[j]) >> 3;
pu1_dst[j * 2] = CLIP_U8(pu1_pred[j * 2] + (pi2_chroma_res[j] + res));
}
pi2_luma_res += luma_res_stride;
pi2_chroma_res += chroma_res_stride;
pu1_dst += dst_stride;
pu1_pred += pred_stride;
}
}

19
common/ihevc_chroma_recon.h
View file

@ -88,13 +88,6 @@ typedef void ihevc_hbd_chroma_recon_16x16_ft(WORD16 *pi2_src,
WORD32 dst_strd,
WORD32 zero_cols,
UWORD8 bit_depth);
typedef void ihevc_chroma_recon_32x32_ft(WORD16 *pi2_src,
UWORD8 *pu1_pred,
UWORD8 *pu1_dst,
WORD32 src_strd,
WORD32 pred_strd,
WORD32 dst_strd,
WORD32 zero_cols);
ihevc_chroma_recon_4x4_ft ihevc_chroma_recon_4x4;
ihevc_hbd_chroma_recon_4x4_ft ihevc_hbd_chroma_recon_4x4;
@ -102,17 +95,5 @@ ihevc_chroma_recon_8x8_ft ihevc_chroma_recon_8x8;
ihevc_hbd_chroma_recon_8x8_ft ihevc_hbd_chroma_recon_8x8;
ihevc_chroma_recon_16x16_ft ihevc_chroma_recon_16x16;
ihevc_hbd_chroma_recon_16x16_ft ihevc_hbd_chroma_recon_16x16;
ihevc_chroma_recon_32x32_ft ihevc_chroma_recon_32x32;
void ihevc_chroma_recon_nxn_ccp(WORD16 *pi2_luma_res,
WORD16 *pi2_chroma_res,
UWORD8 *pu1_pred,
UWORD8 *pu1_dst,
WORD32 alpha,
WORD32 trans_size,
WORD32 luma_res_stride,
WORD32 chroma_res_stride,
WORD32 pred_stride,
WORD32 dst_stride);
#endif /*_IHEVC_CHROMA_RECON_H_*/

3
common/ihevc_defs.h
View file

@ -40,10 +40,7 @@
/*****************************************************************************/
enum
{
IHEVC_PROFILE_UNKNOWN = -1,
IHEVC_PROFILE_MAIN = 0,
IHEVC_PROFILE_MAIN_STILL = 1,
IHEVC_PROFILE_MAIN_REXT = 2,
};
enum

3
common/ihevc_inter_pred.h
View file

@ -46,9 +46,6 @@
#define REF_WIDTH 1280
#define REF_HEIGHT 720
extern WORD8 gai1_ihevc_luma_filter[4][NTAPS_LUMA];
extern WORD8 gai1_ihevc_chroma_filter[8][NTAPS_LUMA];
/*****************************************************************************/
/* Function Declarations */
/*****************************************************************************/

20
common/ihevc_inter_pred_filters.c
View file

@ -61,26 +61,6 @@
#include "ihevc_func_selector.h"
#include "ihevc_inter_pred.h"
WORD8 gai1_ihevc_luma_filter[4][NTAPS_LUMA] =
{
{ 0, 0, 0, 64, 0, 0, 0, 0 },
{ -1, 4, -10, 58, 17, -5, 1, 0 },
{ -1, 4, -11, 40, 40, -11, 4, -1 },
{ 0, 1, -5, 17, 58, -10, 4, -1 } };
/* The filter uses only the first four elements in each array */
WORD8 gai1_ihevc_chroma_filter[8][NTAPS_LUMA] =
{
{ 0, 64, 0, 0, 0, 0, 0, 0 },
{ -2, 58, 10, -2, 0, 0, 0, 0 },
{ -4, 54, 16, -2, 0, 0, 0, 0 },
{ -6, 46, 28, -4, 0, 0, 0, 0 },
{ -4, 36, 36, -4, 0, 0, 0, 0 },
{ -4, 28, 46, -6, 0, 0, 0, 0 },
{ -2, 16, 54, -4, 0, 0, 0, 0 },
{ -2, 10, 58, -2, 0, 0, 0, 0 } };
/*****************************************************************************/
/* Function Definitions */
/*****************************************************************************/

2
common/ihevc_intra_pred.h
View file

@ -159,7 +159,7 @@ typedef void ihevc_intra_pred_ref_filtering_ft(UWORD8 *pu1_src,
WORD32 nt,
UWORD8 *pu1_dst,
WORD32 mode,
WORD32 intra_smoothing_flags);
WORD32 strong_intra_smoothing_enable_flag);
typedef void ihevc_hbd_intra_pred_luma_planar_ft(
UWORD16 *pu2_ref,

31
common/ihevc_intra_pred_filters.c
View file

@ -627,11 +627,6 @@ void ihevc_intra_pred_luma_ref_substitution(UWORD8 *pu1_top_left,
* @param[in] mode
* integer intraprediction mode
*
* @param[in] intra_smoothing_flags
* integer bit 3 indicates if intra smoothing is enabled/disabled
* unconditionally. this is applicable to frext profiles only
* bit 0 indicates strong intra smoothing enabled/disabled
*
* @returns
*
* @remarks
@ -645,7 +640,7 @@ void ihevc_intra_pred_ref_filtering(UWORD8 *pu1_src,
WORD32 nt,
UWORD8 *pu1_dst,
WORD32 mode,
WORD32 intra_smoothing_flags)
WORD32 strong_intra_smoothing_enable_flag)
{
WORD32 filter_flag;
WORD32 i; /* Generic indexing variable */
@ -656,11 +651,9 @@ void ihevc_intra_pred_ref_filtering(UWORD8 *pu1_src,
WORD32 abs_cond_top_flag = 0;
/*WORD32 dc_val = 1 << (BIT_DEPTH - 5);*/
WORD32 dc_val = 1 << (8 - 5);
WORD32 intra_smoothing_disabled = (intra_smoothing_flags >> 3);
WORD32 strong_intra_smoothing_enable_flag = intra_smoothing_flags & 1;
//WORD32 strong_intra_smoothing_enable_flag = 1;
filter_flag = intra_smoothing_disabled ?
0 : (gau1_intra_pred_ref_filter[mode] & (1 << (CTZ(nt) - 2)));
filter_flag = gau1_intra_pred_ref_filter[mode] & (1 << (CTZ(nt) - 2));
if(0 == filter_flag)
{
if(pu1_src == pu1_dst)
@ -950,8 +943,8 @@ void ihevc_intra_pred_luma_dc(UWORD8 *pu1_ref,
* @param[in] nt
* integer Transform Block size
*
* @param[in] disable_boundary_filter
* disable boundary filtering
* @param[in] mode
* integer intraprediction mode
*
* @returns
*
@ -967,16 +960,17 @@ void ihevc_intra_pred_luma_horz(UWORD8 *pu1_ref,
UWORD8 *pu1_dst,
WORD32 dst_strd,
WORD32 nt,
WORD32 disable_boundary_filter)
WORD32 mode)
{
WORD32 row, col;
WORD32 two_nt;
WORD16 s2_predpixel;
UNUSED(mode);
UNUSED(src_strd);
two_nt = 2 * nt;
if(nt == 32 || disable_boundary_filter)
if(nt == 32)
{
for(row = 0; row < nt; row++)
for(col = 0; col < nt; col++)
@ -1029,8 +1023,8 @@ void ihevc_intra_pred_luma_horz(UWORD8 *pu1_ref,
* @param[in] nt
* integer Transform Block size
*
* @param[in] disable_boundary_filter
* disable boundary filtering
* @param[in] mode
* integer intraprediction mode
*
* @returns
*
@ -1046,14 +1040,15 @@ void ihevc_intra_pred_luma_ver(UWORD8 *pu1_ref,
UWORD8 *pu1_dst,
WORD32 dst_strd,
WORD32 nt,
WORD32 disable_boundary_filter)
WORD32 mode)
{
WORD32 row, col;
WORD16 s2_predpixel;
WORD32 two_nt = 2 * nt;
UNUSED(mode);
UNUSED(src_strd);
if(nt == 32 || disable_boundary_filter)
if(nt == 32)
{
/* Replication to next columns*/
for(row = 0; row < nt; row++)

2493
common/ihevc_itrans_res.c
View file

File diff suppressed because it is too large Load diff

108
common/ihevc_itrans_res.h
View file

@ -1,108 +0,0 @@
/******************************************************************************
*
* 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_res.h
*
* @brief
* Functions declarations for inverse transform
*
* @author
* Ittiam
*
* @remarks
* None
*
*******************************************************************************
*/
#ifndef _IHEVC_ITRANS_RES_H_
#define _IHEVC_ITRANS_RES_H_
typedef void ihevc_itrans_res_4x4_ttype1_ft(WORD16 *pi2_src,
WORD16 *pi2_tmp,
WORD16 *pi2_dst,
WORD32 src_strd,
WORD32 dst_strd,
WORD32 zero_cols,
WORD32 zero_rows);
typedef void ihevc_itrans_res_4x4_ft(WORD16 *pi2_src,
WORD16 *pi2_tmp,
WORD16 *pi2_dst,
WORD32 src_strd,
WORD32 dst_strd,
WORD32 zero_cols,
WORD32 zero_rows);
typedef void ihevcd_itrans_res_dc_ft(WORD16 *pi2_dst,
WORD32 dst_strd,
WORD32 log2_trans_size,
WORD16 i2_coeff_value);
typedef void ihevc_itrans_res_8x8_ft(WORD16 *pi2_src,
WORD16 *pi2_tmp,
WORD16 *pi2_dst,
WORD32 src_strd,
WORD32 dst_strd,
WORD32 zero_cols,
WORD32 zero_rows);
typedef void ihevc_itrans_res_16x16_ft(WORD16 *pi2_src,
WORD16 *pi2_tmp,
WORD16 *pi2_dst,
WORD32 src_strd,
WORD32 dst_strd,
WORD32 zero_cols,
WORD32 zero_rows);
typedef void ihevc_itrans_res_32x32_ft(WORD16 *pi2_src,
WORD16 *pi2_tmp,
WORD16 *pi2_dst,
WORD32 src_strd,
WORD32 dst_strd,
WORD32 zero_cols,
WORD32 zero_rows);
typedef void ihevc_res_4x4_transform(WORD16 *pi2_src,
WORD16 *pi2_dst,
WORD32 src_strd,
WORD32 dst_strd,
WORD32 zero_cols);
typedef void ihevc_res_nxn_transform(WORD16 *pi2_src,
WORD16 *pi2_dst,
WORD32 src_strd,
WORD32 dst_strd,
WORD32 trans_size,
WORD32 zero_cols);
/* C function declarations */
ihevc_itrans_res_4x4_ttype1_ft ihevc_itrans_res_4x4_ttype1;
ihevc_itrans_res_4x4_ft ihevc_itrans_res_4x4;
ihevcd_itrans_res_dc_ft ihevcd_itrans_res_dc;
ihevc_itrans_res_8x8_ft ihevc_itrans_res_8x8;
ihevc_itrans_res_16x16_ft ihevc_itrans_res_16x16;
ihevc_itrans_res_32x32_ft ihevc_itrans_res_32x32;
ihevc_res_4x4_transform ihevc_res_4x4_rotate;
ihevc_res_nxn_transform ihevc_res_nxn_copy;
ihevc_res_nxn_transform ihevc_res_nxn_rdpcm_horz;
ihevc_res_nxn_transform ihevc_res_nxn_rdpcm_vert;
#endif /*_IHEVC_ITRANS_RES_H_*/

96
common/ihevc_structs.h
View file

@ -1420,28 +1420,6 @@ typedef struct
*/
UWORD32 b3_chroma_intra_mode_idx : 3;
#ifdef ENABLE_MAIN_REXT_PROFILE
/**
* Cb CCP alpha magnitude
*/
UWORD32 b3_cb_log2_res_scale_abs_plus1 : 3;
/**
* Cb CCP alpha sign
*/
UWORD32 b1_cb_log2_res_sign : 1;
/**
* Cr CCP alpha magnitude
*/
UWORD32 b3_cr_log2_res_scale_abs_plus1 : 3;
/**
* Cr CCP alpha sign
*/
UWORD32 b1_cr_log2_res_sign : 1;
#endif
}tu_t;
/**
@ -2503,36 +2481,6 @@ typedef struct
*/
vui_t s_vui_parameters;
/**
* sps_extension_present_flag
*/
WORD8 i1_sps_extension_present_flag;
/**
* sps_range_extension_present_flag
*/
WORD8 i1_sps_range_extension_flag;
/**
* sps_multilayer_extension_present_flag
*/
WORD8 i1_sps_multilayer_extension_flag;
/**
* sps_3d_extension_present_flag
*/
WORD8 i1_sps_3d_extension_flag;
/**
* sps_scc_extension_present_flag
*/
WORD8 i1_sps_scc_extension_flag;
/**
* sps_extension_reserved
*/
WORD8 i1_sps_extension_4bits;
/**
* Log2(CTB Size) in luma units
*/
@ -2595,8 +2543,8 @@ typedef struct
/* Inter 8 x 8 Y, 8 x 8 U, 8 x 8 V */
/* Intra 16x16 Y, 16x16 U, 16x16 V */
/* Inter 16x16 Y, 16x16 U, 16x16 V */
/* Intra 32x32 Y, 32x32 U, 32x32 V */
/* Inter 32x32 Y, 32x32 U, 32x32 V */
/* Intra 32x32 Y */
/* Inter 32x32 Y */
/*************************************************************************/
WORD16 *pi2_scaling_mat;
@ -2638,9 +2586,9 @@ typedef struct
WORD8 i1_use_high_precision_pred_wt;
/**
* persistent_rice_adaptation_enabled_flag
* fast_rice_adaptation_enabled_flag
*/
WORD8 i1_persistent_rice_adaptation_enabled_flag;
WORD8 i1_fast_rice_adaptation_enabled_flag;
/**
* cabac_bypass_alignment_enabled_flag
@ -2671,8 +2619,8 @@ typedef struct
/* Inter 8 x 8 Y, 8 x 8 U, 8 x 8 V */
/* Intra 16x16 Y, 16x16 U, 16x16 V */
/* Inter 16x16 Y, 16x16 U, 16x16 V */
/* Intra 32x32 Y, 32x32 U, 32x32 V */
/* Inter 32x32 Y, 32x32 U, 32x32 V */
/* Intra 32x32 Y */
/* Inter 32x32 Y */
/*************************************************************************/
WORD16 *pi2_scaling_mat;
@ -2874,36 +2822,6 @@ typedef struct
*/
WORD8 i1_slice_extension_present_flag;
/**
* pps_extension_present_flag
*/
WORD8 i1_pps_extension_present_flag;
/**
* pps_range_extension_present_flag
*/
WORD8 i1_pps_range_extension_flag;
/**
* pps_multilayer_extension_present_flag
*/
WORD8 i1_pps_multilayer_extension_flag;
/**
* pps_3d_extension_present_flag
*/
WORD8 i1_pps_3d_extension_flag;
/**
* pps_scc_extension_present_flag
*/
WORD8 i1_pps_scc_extension_flag;
/**
* pps_extension_reserved
*/
WORD8 i1_pps_extension_4bits;
/**
* scaling_list_dc_coef_minus8
*/
@ -2924,7 +2842,7 @@ typedef struct
/**
* log2_max_transform_skip_block_size_minus2
*/
WORD8 i1_log2_max_transform_skip_block_size_minus2;
WORD32 i4_log2_max_transform_skip_block_size_minus2;
/**
* cross_component_prediction_enabled_flag

6
common/ihevc_trans_tables.c
View file

@ -877,12 +877,6 @@ const WORD16 g_ai2_ihevc_trans_intr_4[4][8] =
const UWORD8 IHEVCE_CHROMA_SHUFFLEMASK_HBD[8] = { 0x00, 0x01, 0x04, 0x05,
0x08, 0x09, 0x0C, 0x0D };
const UWORD32 gau4_ihevcd_4_bit_reverse[] =
{
0, 8, 4, 12, 2, 10, 6, 14, 1, 9, 5, 13, 3, 11, 7, 15
};
#ifndef DISABLE_AVX2
const WORD32 g_ai4_ihevc_trans_8_intr_avx2[7][8] =
{ /* 4*32 = 128 bit */

2
common/ihevc_trans_tables.h
View file

@ -116,6 +116,4 @@ extern const WORD16 g_ai2_ihevc_trans_intr_4[4][8];
extern const UWORD8 IHEVCE_CHROMA_SHUFFLEMASK_HBD[8];
extern MEM_ALIGN16 const UWORD32 gau4_ihevcd_4_bit_reverse[16];
#endif /*_IHEVC_TRANS_TABLES_H_*/

16
common/x86/ihevc_intra_pred_filters_sse42_intr.c
View file

@ -91,11 +91,6 @@
* @param[in] mode
* integer intraprediction mode
*
* @param[in] intra_smoothing_flags
* integer bit 3 indicates if intra smoothing is enabled/disabled
* unconditionally. this is applicable to frext profiles only
* bit 0 indicates strong intra smoothing enabled/disabled
*
* @returns
*
* @remarks
@ -109,7 +104,7 @@ void ihevc_intra_pred_ref_filtering_sse42(UWORD8 *pu1_src,
WORD32 nt,
UWORD8 *pu1_dst,
WORD32 mode,
WORD32 intra_smoothing_flags)
WORD32 strong_intra_smoothing_enable_flag)
{
WORD32 filter_flag;
WORD32 i; /* Generic indexing variable */
@ -122,10 +117,11 @@ void ihevc_intra_pred_ref_filtering_sse42(UWORD8 *pu1_src,
__m128i src_temp1, src_temp2, src_temp3, src_temp7;
__m128i src_temp4, src_temp5, src_temp6, src_temp8;
WORD32 intra_smoothing_disabled = (intra_smoothing_flags >> 3);
WORD32 strong_intra_smoothing_enable_flag = intra_smoothing_flags & 1;
filter_flag = intra_smoothing_disabled ?
0 : (gau1_intra_pred_ref_filter[mode] & (1 << (CTZ(nt) - 2)));
//WORD32 strong_intra_smoothing_enable_flag = 1;
filter_flag = gau1_intra_pred_ref_filter[mode] & (1 << (CTZ(nt) - 2));
if(0 == filter_flag)
{
if(pu1_src == pu1_dst)

14
common/x86/ihevc_intra_pred_filters_ssse3_intr.c
View file

@ -377,11 +377,6 @@ void ihevc_intra_pred_luma_ref_substitution_ssse3(UWORD8 *pu1_top_left,
* @param[in] mode
* integer intraprediction mode
*
* @param[in] intra_smoothing_flags
* integer bit 3 indicates if intra smoothing is enabled/disabled
* unconditionally. this is applicable to frext profiles only
* bit 0 indicates strong intra smoothing enabled/disabled
*
* @returns
*
* @remarks
@ -394,7 +389,7 @@ void ihevc_intra_pred_ref_filtering_ssse3(UWORD8 *pu1_src,
WORD32 nt,
UWORD8 *pu1_dst,
WORD32 mode,
WORD32 intra_smoothing_flags)
WORD32 strong_intra_smoothing_enable_flag)
{
WORD32 filter_flag;
WORD32 i; /* Generic indexing variable */
@ -407,10 +402,9 @@ void ihevc_intra_pred_ref_filtering_ssse3(UWORD8 *pu1_src,
__m128i src_temp1, src_temp2, src_temp3, src_temp7;
__m128i src_temp4, src_temp5, src_temp6, src_temp8;
WORD32 intra_smoothing_disabled = (intra_smoothing_flags >> 3);
WORD32 strong_intra_smoothing_enable_flag = intra_smoothing_flags & 1;
filter_flag = intra_smoothing_disabled ?
0 : (gau1_intra_pred_ref_filter[mode] & (1 << (CTZ(nt) - 2)));
//WORD32 strong_intra_smoothing_enable_flag = 1;
filter_flag = gau1_intra_pred_ref_filter[mode] & (1 << (CTZ(nt) - 2));
if(0 == filter_flag)
{
if(pu1_src == pu1_dst)

453
decoder/arm/ihevcd_fmt_conv_420sp_to_rgba8888.s Normal file
View file

@ -0,0 +1,453 @@
@/*****************************************************************************
@*
@* 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
@* ihevcd_fmt_conv_420sp_to_rgba8888.s
@*
@* @brief
@* contains function definitions for format conversions
@*
@* @author
@* ittiam
@*
@* @par list of functions:
@*
@*
@* @remarks
@* none
@*
@*******************************************************************************/
.equ DO1STROUNDING, 0
@ ARM
@
@ PRESERVE8
.text
.p2align 2
@/*****************************************************************************
@* *
@* Function Name : ihevcd_fmt_conv_420sp_to_rgba8888() *
@* *
@* Description : This function conversts the image from YUV422 color *
@* space to RGB888 color space. The function can be *
@* invoked at the MB level. *
@* *
@* Arguments : R0 pubY *
@* R1 pubUV *
@* R2 pusRGB *
@* R3 pusRGB *
@* [R13 #40] usHeight *
@* [R13 #44] usWidth *
@* [R13 #48] usStrideY *
@* [R13 #52] usStrideU *
@* [R13 #56] usStrideV *
@* [R13 #60] usStrideRGB *
@* *
@* Values Returned : None *
@* *
@* Register Usage : R0 - R14 *
@* *
@* Stack Usage : 104 Bytes *
@* *
@* Interruptibility : Interruptible *
@* *
@* Known Limitations *
@* Assumptions: Image Width: Assumed to be multiple of 16 and *
@* greater than or equal to 16 *
@* Image Height: Assumed to be even. *
@* *
@* Revision History : *
@* DD MM YYYY Author(s) Changes (Describe the changes made) *
@* 07 06 2010 Varshita Draft *
@* 07 06 2010 Naveen Kr T Completed *
@* 05 08 2013 Naveen K P Modified for HEVC *
@* 30 10 2018 Saurabh Sood Store D registers to stack *
@*****************************************************************************/
.global ihevcd_fmt_conv_420sp_to_rgba8888_a9q
.type ihevcd_fmt_conv_420sp_to_rgba8888_a9q, function
ihevcd_fmt_conv_420sp_to_rgba8888_a9q:
@// push the registers on the stack
STMFD SP!,{R4-R12,LR}
VPUSH {d8-d15}
@//R0 - Y PTR
@//R1 - UV PTR
@//R2 - RGB PTR
@//R3 - RGB PTR
@//R4 - PIC WIDTH
@//R5 - PIC HT
@//R6 - STRIDE Y
@//R7 - STRIDE U
@//R8 - STRIDE V
@//R9 - STRIDE RGB
@//ONE ROW PROCESSING AT A TIME
@//THE FOUR CONSTANTS ARE:
@//C1=0x3311,C2=0xF379,C3=0xE5F8,C4=0x4092
@PLD [R0]
@PLD [R1]
@PLD [R2]
@/* can be loaded from a defined const type */
MOVW R10,#0x3311
VMOV.16 D0[0],R10 @//C1
MOVW R10,#0xF379
VMOV.16 D0[1],R10 @//C2
MOVW R10,#0xE5F8
VMOV.16 D0[2],R10 @//C3
MOVW R10,#0x4092
VMOV.16 D0[3],R10 @//C4
@//LOAD CONSTANT 128 INTO A CORTEX REGISTER
MOV R10,#128
VDUP.8 D1,R10
@//D0 HAS C1-C2-C3-C4
@// load other parameters from stack
LDR R5,[sp,#104]
@LDR R4,[sp,#44]
LDR R6,[sp,#108]
LDR R7,[sp,#112]
@LDR R8,[sp,#52]
LDR R9,[sp,#116]
@// calculate offsets, offset = stride - width
SUB R10,R6,R3 @// luma offset
SUB R11,R7,R3
@, LSR #1 @// u offset
@SUB R12,R8,R3, LSR #1 @// v offset
SUB R14,R9,R3 @// rgb offset in pixels
@// calculate height loop count
MOV R5,R5, LSR #1 @// height_cnt = height / 16
@// create next row pointers for rgb and luma data
ADD R7,R0,R6 @// luma_next_row = luma + luma_stride
ADD R8,R2,R9,LSL #2 @// rgb_next_row = rgb + rgb_stride
LABEL_YUV420SP_TO_RGB8888_HEIGHT_LOOP:
@//LOAD VALUES OF U&V AND COMPUTE THE R,G,B WEIGHT VALUES.
VLD1.8 {D2,D3},[R1]! @//LOAD 8 VALUES OF UV
@//VLD1.8 {D3},[R2]! @//LOAD 8 VALUES OF V
@// calculate width loop count
MOV R6,R3, LSR #4 @// width_cnt = width / 16
@//COMPUTE THE ACTUAL RGB VALUES,WE CAN DO TWO ROWS AT A TIME
@//LOAD VALUES OF Y 8-BIT VALUES
VLD2.8 {D30,D31},[R0]! @//D0 - Y0,Y2,Y4,Y6,Y8,Y10,Y12,Y14 row 1
@//D1 - Y1,Y3,Y5,Y7,Y9,Y11,Y13,Y15
VLD2.8 {D28,D29},[R7]! @//D0 - Y0,Y2,Y4,Y6,Y8,Y10,Y12,Y14 row2
@//D1 - Y1,Y3,Y5,Y7,Y9,Y11,Y13,Y15
SUBS R6,R6,#1
BEQ LABEL_YUV420SP_TO_RGB8888_WIDTH_LOOP_SKIP
LABEL_YUV420SP_TO_RGB8888_WIDTH_LOOP:
@VMOV.I8 Q1,#128
VUZP.8 D2,D3
@//NEED TO SUBTRACT (U-128) AND (V-128)
@//(D2-D1),(D3-D1)
VSUBL.U8 Q2,D2,D1 @//(U-128)
VSUBL.U8 Q3,D3,D1 @//(V-128)
@//LOAD VALUES OF U&V for next row
VLD1.8 {D2,D3},[R1]! @//LOAD 8 VALUES OF U
@//VLD1.8 {D3},[R2]! @//LOAD 8 VALUES OF V
@PLD [R0]
PLD [R1]
@//NEED TO MULTIPLY WITH Q2,Q3 WITH CO-EEFICIENTS
VMULL.S16 Q4,D4,D0[3] @//(U-128)*C4 FOR B
VMULL.S16 Q5,D5,D0[3] @//(U-128)*C4 FOR B
VMULL.S16 Q10,D6,D0[0] @//(V-128)*C1 FOR R
VMULL.S16 Q11,D7,D0[0] @//(V-128)*C1 FOR R
VMULL.S16 Q6,D4,D0[1] @//(U-128)*C2 FOR G
VMLAL.S16 Q6,D6,D0[2] @//Q6 = (U-128)*C2 + (V-128)*C3
VMULL.S16 Q7,D5,D0[1] @//(U-128)*C2 FOR G
VMLAL.S16 Q7,D7,D0[2] @//Q7 = (U-128)*C2 + (V-128)*C3
@//NARROW RIGHT SHIFT BY 13 FOR R&B
VQSHRN.S32 D8,Q4,#13 @//D8 = (U-128)*C4>>13 4 16-BIT VALUES
VQSHRN.S32 D9,Q5,#13 @//D9 = (U-128)*C4>>13 4 16-BIT VALUES
@//Q4 - WEIGHT FOR B
@//NARROW RIGHT SHIFT BY 13 FOR R&B
VQSHRN.S32 D10,Q10,#13 @//D10 = (V-128)*C1>>13 4 16-BIT VALUES
VQSHRN.S32 D11,Q11,#13 @//D11 = (V-128)*C1>>13 4 16-BIT VALUES
@//Q5 - WEIGHT FOR R
@//NARROW RIGHT SHIFT BY 13 FOR G
VQSHRN.S32 D12,Q6,#13 @//D12 = [(U-128)*C2 + (V-128)*C3]>>13 4 16-BIT VALUES
VQSHRN.S32 D13,Q7,#13 @//D13 = [(U-128)*C2 + (V-128)*C3]>>13 4 16-BIT VALUES
@//Q6 - WEIGHT FOR G
VADDW.U8 Q7,Q4,D30 @//Q7 - HAS Y + B
VADDW.U8 Q8,Q5,D30 @//Q8 - HAS Y + R
VADDW.U8 Q9,Q6,D30 @//Q9 - HAS Y + G
VADDW.U8 Q10,Q4,D31 @//Q10 - HAS Y + B
VADDW.U8 Q11,Q5,D31 @//Q11 - HAS Y + R
VADDW.U8 Q12,Q6,D31 @//Q12 - HAS Y + G
VQMOVUN.S16 D14,Q7
VQMOVUN.S16 D15,Q9
VQMOVUN.S16 D16,Q8
VMOV.I8 D17,#0
VZIP.8 D14,D15
VZIP.8 D16,D17
VZIP.16 Q7,Q8
VQMOVUN.S16 D20,Q10
VQMOVUN.S16 D21,Q12
VQMOVUN.S16 D22,Q11
VMOV.I8 D23,#0
VZIP.8 D20,D21
VZIP.8 D22,D23
VZIP.16 Q10,Q11
VZIP.32 Q7,Q10
VZIP.32 Q8,Q11
VST1.32 D14,[R2]!
VST1.32 D15,[R2]!
VST1.32 D20,[R2]!
VST1.32 D21,[R2]!
VST1.32 D16,[R2]!
VST1.32 D17,[R2]!
VST1.32 D22,[R2]!
VST1.32 D23,[R2]!
@//D14-D20 - TOALLY HAVE 16 VALUES
@//WE NEED TO SHIFT R,G,B VALUES TO GET 5BIT,6BIT AND 5BIT COMBINATIONS
VADDW.U8 Q7,Q4,D28 @//Q7 - HAS Y + B
VADDW.U8 Q8,Q5,D28 @//Q2 - HAS Y + R
VADDW.U8 Q9,Q6,D28 @//Q3 - HAS Y + G
VADDW.U8 Q10,Q4,D29 @//Q10 - HAS Y + B
VADDW.U8 Q11,Q5,D29 @//Q11 - HAS Y + R
VADDW.U8 Q12,Q6,D29 @//Q12 - HAS Y + G
@//COMPUTE THE ACTUAL RGB VALUES,WE CAN DO TWO ROWS AT A TIME
@//LOAD VALUES OF Y 8-BIT VALUES
VLD2.8 {D30,D31},[R0]! @//D0 - Y0,Y2,Y4,Y6,Y8,Y10,Y12,Y14 row 1
@//D1 - Y1,Y3,Y5,Y7,Y9,Y11,Y13,Y15
VLD2.8 {D28,D29},[R7]! @//D0 - Y0,Y2,Y4,Y6,Y8,Y10,Y12,Y14 row2
@//D1 - Y1,Y3,Y5,Y7,Y9,Y11,Y13,Y15
PLD [R0]
PLD [R7]
VQMOVUN.S16 D14,Q7
VQMOVUN.S16 D15,Q9
VQMOVUN.S16 D16,Q8
VMOV.I8 D17,#0
VZIP.8 D14,D15
VZIP.8 D16,D17
VZIP.16 Q7,Q8
VQMOVUN.S16 D20,Q10
VQMOVUN.S16 D21,Q12
VQMOVUN.S16 D22,Q11
VMOV.I8 D23,#0
VZIP.8 D20,D21
VZIP.8 D22,D23
VZIP.16 Q10,Q11
VZIP.32 Q7,Q10
VZIP.32 Q8,Q11
VST1.32 D14,[R8]!
VST1.32 D15,[R8]!
VST1.32 D20,[R8]!
VST1.32 D21,[R8]!
VST1.32 D16,[R8]!
VST1.32 D17,[R8]!
VST1.32 D22,[R8]!
VST1.32 D23,[R8]!
SUBS R6,R6,#1 @// width_cnt -= 1
BNE LABEL_YUV420SP_TO_RGB8888_WIDTH_LOOP
LABEL_YUV420SP_TO_RGB8888_WIDTH_LOOP_SKIP:
@VMOV.I8 Q1,#128
VUZP.8 D2,D3
@//NEED TO SUBTRACT (U-128) AND (V-128)
@//(D2-D1),(D3-D1)
VSUBL.U8 Q2,D2,D1 @//(U-128)
VSUBL.U8 Q3,D3,D1 @//(V-128)
@//NEED TO MULTIPLY WITH Q2,Q3 WITH CO-EEFICIENTS
VMULL.S16 Q4,D4,D0[3] @//(U-128)*C4 FOR B
VMULL.S16 Q5,D5,D0[3] @//(U-128)*C4 FOR B
VMULL.S16 Q10,D6,D0[0] @//(V-128)*C1 FOR R
VMULL.S16 Q11,D7,D0[0] @//(V-128)*C1 FOR R
VMULL.S16 Q6,D4,D0[1] @//(U-128)*C2 FOR G
VMLAL.S16 Q6,D6,D0[2] @//Q6 = (U-128)*C2 + (V-128)*C3
VMULL.S16 Q7,D5,D0[1] @//(U-128)*C2 FOR G
VMLAL.S16 Q7,D7,D0[2] @//Q7 = (U-128)*C2 + (V-128)*C3
@//NARROW RIGHT SHIFT BY 13 FOR R&B
VQSHRN.S32 D8,Q4,#13 @//D8 = (U-128)*C4>>13 4 16-BIT VALUES
VQSHRN.S32 D9,Q5,#13 @//D9 = (U-128)*C4>>13 4 16-BIT VALUES
@//Q4 - WEIGHT FOR B
@//NARROW RIGHT SHIFT BY 13 FOR R&B
VQSHRN.S32 D10,Q10,#13 @//D10 = (V-128)*C1>>13 4 16-BIT VALUES
VQSHRN.S32 D11,Q11,#13 @//D11 = (V-128)*C1>>13 4 16-BIT VALUES
@//Q5 - WEIGHT FOR R
@//NARROW RIGHT SHIFT BY 13 FOR G
VQSHRN.S32 D12,Q6,#13 @//D12 = [(U-128)*C2 + (V-128)*C3]>>13 4 16-BIT VALUES
VQSHRN.S32 D13,Q7,#13 @//D13 = [(U-128)*C2 + (V-128)*C3]>>13 4 16-BIT VALUES
@//Q6 - WEIGHT FOR G
VADDW.U8 Q7,Q4,D30 @//Q7 - HAS Y + B
VADDW.U8 Q8,Q5,D30 @//Q8 - HAS Y + R
VADDW.U8 Q9,Q6,D30 @//Q9 - HAS Y + G
VADDW.U8 Q10,Q4,D31 @//Q10 - HAS Y + B
VADDW.U8 Q11,Q5,D31 @//Q11 - HAS Y + R
VADDW.U8 Q12,Q6,D31 @//Q12 - HAS Y + G
VQMOVUN.S16 D14,Q7
VQMOVUN.S16 D15,Q9
VQMOVUN.S16 D16,Q8
VMOV.I8 D17,#0
VZIP.8 D14,D15
VZIP.8 D16,D17
VZIP.16 Q7,Q8
VQMOVUN.S16 D20,Q10
VQMOVUN.S16 D21,Q12
VQMOVUN.S16 D22,Q11
VMOV.I8 D23,#0
VZIP.8 D20,D21
VZIP.8 D22,D23
VZIP.16 Q10,Q11
VZIP.32 Q7,Q10
VZIP.32 Q8,Q11
VST1.32 D14,[R2]!
VST1.32 D15,[R2]!
VST1.32 D20,[R2]!
VST1.32 D21,[R2]!
VST1.32 D16,[R2]!
VST1.32 D17,[R2]!
VST1.32 D22,[R2]!
VST1.32 D23,[R2]!
@//D14-D20 - TOALLY HAVE 16 VALUES
@//WE NEED TO SHIFT R,G,B VALUES TO GET 5BIT,6BIT AND 5BIT COMBINATIONS
VADDW.U8 Q7,Q4,D28 @//Q7 - HAS Y + B
VADDW.U8 Q8,Q5,D28 @//Q2 - HAS Y + R
VADDW.U8 Q9,Q6,D28 @//Q3 - HAS Y + G
VADDW.U8 Q10,Q4,D29 @//Q10 - HAS Y + B
VADDW.U8 Q11,Q5,D29 @//Q11 - HAS Y + R
VADDW.U8 Q12,Q6,D29 @//Q12 - HAS Y + G
VQMOVUN.S16 D14,Q7
VQMOVUN.S16 D15,Q9
VQMOVUN.S16 D16,Q8
VMOV.I8 D17,#0
VZIP.8 D14,D15
VZIP.8 D16,D17
VZIP.16 Q7,Q8
VQMOVUN.S16 D20,Q10
VQMOVUN.S16 D21,Q12
VQMOVUN.S16 D22,Q11
VMOV.I8 D23,#0
VZIP.8 D20,D21
VZIP.8 D22,D23
VZIP.16 Q10,Q11
VZIP.32 Q7,Q10
VZIP.32 Q8,Q11
VST1.32 D14,[R8]!
VST1.32 D15,[R8]!
VST1.32 D20,[R8]!
VST1.32 D21,[R8]!
VST1.32 D16,[R8]!
VST1.32 D17,[R8]!
VST1.32 D22,[R8]!
VST1.32 D23,[R8]!
@// Adjust the address pointers
ADD R0,R7,R10 @// luma = luma_next + offset
ADD R2,R8,R14,LSL #2 @// rgb = rgb_next + offset
ADD R7,R0,R3 @// luma_next = luma + width
ADD R8,R2,R3,LSL #2 @// rgb_next_row = rgb + width
ADD R1,R1,R11 @// adjust u pointer
@ADD R2,R2,R12 @// adjust v pointer
ADD R7,R7,R10 @// luma_next = luma + width + offset (because of register crunch)
ADD R8,R8,R14,LSL #2 @// rgb_next_row = rgb + width + offset
SUBS R5,R5,#1 @// height_cnt -= 1
BNE LABEL_YUV420SP_TO_RGB8888_HEIGHT_LOOP
@//POP THE REGISTERS
VPOP {d8-d15}
LDMFD SP!,{R4-R12,PC}
.section .note.GNU-stack,"",%progbits

21
decoder/arm/ihevcd_function_selector.c
View file

@ -58,6 +58,10 @@
#include "ihevcd_function_selector.h"
#include "ihevcd_structs.h"
void ihevcd_init_function_ptr_neonintr(codec_t *ps_codec);
void ihevcd_init_function_ptr_noneon(codec_t *ps_codec);
void ihevcd_init_function_ptr_a9q(codec_t *ps_codec);
void ihevcd_init_function_ptr_av8(codec_t *ps_codec);
void ihevcd_init_function_ptr(void *pv_codec)
{
codec_t *ps_codec = (codec_t *)pv_codec;
@ -67,11 +71,11 @@ void ihevcd_init_function_ptr(void *pv_codec)
{
#ifndef DISABLE_NEONINTR
case ARCH_ARM_NEONINTR:
ihevcd_init_function_ptr_neonintr(&ps_codec->s_func_selector);
ihevcd_init_function_ptr_neonintr(ps_codec);
break;
#endif
case ARCH_ARM_NONEON:
ihevcd_init_function_ptr_noneon(&ps_codec->s_func_selector);
ihevcd_init_function_ptr_noneon(ps_codec);
break;
default:
case ARCH_ARM_A5:
@ -80,9 +84,9 @@ void ihevcd_init_function_ptr(void *pv_codec)
case ARCH_ARM_A15:
case ARCH_ARM_A9Q:
#ifndef DISABLE_NEON
ihevcd_init_function_ptr_a9q(&ps_codec->s_func_selector);
ihevcd_init_function_ptr_a9q(ps_codec);
#else
ihevcd_init_function_ptr_noneon(&ps_codec->s_func_selector);
ihevcd_init_function_ptr_noneon(ps_codec);
#endif
break;
}
@ -102,17 +106,12 @@ void ihevcd_init_function_ptr(void *pv_codec)
switch(ps_codec->e_processor_arch)
{
case ARCH_ARM_NONEON:
ihevcd_init_function_ptr_noneon(&ps_codec->s_func_selector);
ihevcd_init_function_ptr_noneon(ps_codec);
break;
case ARCH_ARMV8_GENERIC:
default:
#ifdef DARWIN
ihevcd_init_function_ptr_noneon(&ps_codec->s_func_selector);
ihevcd_init_function_ptr_av8(ps_codec);
break;
#else
ihevcd_init_function_ptr_av8(&ps_codec->s_func_selector);
break;
#endif
}
#endif
}

206
decoder/arm/ihevcd_function_selector_a9q.c
View file

@ -54,113 +54,107 @@
#include "ihevc_dpb_mgr.h"
#include "ihevc_error.h"
#include "ihevcd_defs.h"
#include "ihevcd_function_selector.h"
#include "ihevcd_structs.h"
void ihevcd_init_function_ptr_a9q(func_selector_t *ps_func_selector)
void ihevcd_init_function_ptr_a9q(codec_t *ps_codec)
{
ps_func_selector->ihevc_deblk_chroma_horz_fptr = &ihevc_deblk_chroma_horz_a9q;
ps_func_selector->ihevc_deblk_chroma_vert_fptr = &ihevc_deblk_chroma_vert_a9q;
ps_func_selector->ihevc_deblk_luma_vert_fptr = &ihevc_deblk_luma_vert_a9q;
ps_func_selector->ihevc_deblk_luma_horz_fptr = &ihevc_deblk_luma_horz_a9q;
ps_func_selector->ihevc_inter_pred_chroma_copy_fptr = &ihevc_inter_pred_chroma_copy_a9q;
ps_func_selector->ihevc_inter_pred_chroma_copy_w16out_fptr = &ihevc_inter_pred_chroma_copy_w16out_a9q;
ps_func_selector->ihevc_inter_pred_chroma_horz_fptr = &ihevc_inter_pred_chroma_horz_a9q;
ps_func_selector->ihevc_inter_pred_chroma_horz_w16out_fptr = &ihevc_inter_pred_chroma_horz_w16out_a9q;
ps_func_selector->ihevc_inter_pred_chroma_vert_fptr = &ihevc_inter_pred_chroma_vert_a9q;
ps_func_selector->ihevc_inter_pred_chroma_vert_w16inp_fptr = &ihevc_inter_pred_chroma_vert_w16inp_a9q;
ps_func_selector->ihevc_inter_pred_chroma_vert_w16inp_w16out_fptr = &ihevc_inter_pred_chroma_vert_w16inp_w16out_a9q;
ps_func_selector->ihevc_inter_pred_chroma_vert_w16out_fptr = &ihevc_inter_pred_chroma_vert_w16out_a9q;
ps_func_selector->ihevc_inter_pred_luma_horz_fptr = &ihevc_inter_pred_luma_horz_a9q;
ps_func_selector->ihevc_inter_pred_luma_vert_fptr = &ihevc_inter_pred_luma_vert_a9q;
ps_func_selector->ihevc_inter_pred_luma_vert_w16out_fptr = &ihevc_inter_pred_luma_vert_w16out_a9q;
ps_func_selector->ihevc_inter_pred_luma_vert_w16inp_fptr = &ihevc_inter_pred_luma_vert_w16inp_a9q;
ps_func_selector->ihevc_inter_pred_luma_copy_fptr = &ihevc_inter_pred_luma_copy_a9q;
ps_func_selector->ihevc_inter_pred_luma_copy_w16out_fptr = &ihevc_inter_pred_luma_copy_w16out_a9q;
ps_func_selector->ihevc_inter_pred_luma_horz_w16out_fptr = &ihevc_inter_pred_luma_horz_w16out_a9q;
ps_func_selector->ihevc_inter_pred_luma_vert_w16inp_w16out_fptr = &ihevc_inter_pred_luma_vert_w16inp_w16out_a9q;
ps_func_selector->ihevc_intra_pred_chroma_ref_substitution_fptr = &ihevc_intra_pred_chroma_ref_substitution_a9q;
ps_func_selector->ihevc_intra_pred_luma_ref_substitution_fptr = &ihevc_intra_pred_luma_ref_substitution_a9q;
ps_func_selector->ihevc_intra_pred_luma_ref_subst_all_avlble_fptr = &ihevc_intra_pred_luma_ref_subst_all_avlble;
ps_func_selector->ihevc_intra_pred_ref_filtering_fptr = &ihevc_intra_pred_ref_filtering_neonintr;
ps_func_selector->ihevc_intra_pred_chroma_ref_filtering_fptr = &ihevc_intra_pred_chroma_ref_filtering;
ps_func_selector->ihevc_intra_pred_chroma_dc_fptr = &ihevc_intra_pred_chroma_dc_a9q;
ps_func_selector->ihevc_intra_pred_chroma_horz_fptr = &ihevc_intra_pred_chroma_horz_a9q;
ps_func_selector->ihevc_intra_pred_chroma_mode2_fptr = &ihevc_intra_pred_chroma_mode2_a9q;
ps_func_selector->ihevc_intra_pred_chroma_mode_18_34_fptr = &ihevc_intra_pred_chroma_mode_18_34_a9q;
ps_func_selector->ihevc_intra_pred_chroma_mode_27_to_33_fptr = &ihevc_intra_pred_chroma_mode_27_to_33_a9q;
ps_func_selector->ihevc_intra_pred_chroma_mode_3_to_9_fptr = &ihevc_intra_pred_chroma_mode_3_to_9_a9q;
ps_func_selector->ihevc_intra_pred_chroma_planar_fptr = &ihevc_intra_pred_chroma_planar_a9q;
ps_func_selector->ihevc_intra_pred_chroma_ver_fptr = &ihevc_intra_pred_chroma_ver_a9q;
ps_func_selector->ihevc_intra_pred_chroma_mode_11_to_17_fptr = &ihevc_intra_pred_chroma_mode_11_to_17_a9q;
ps_func_selector->ihevc_intra_pred_chroma_mode_19_to_25_fptr = &ihevc_intra_pred_chroma_mode_19_to_25_a9q;
ps_func_selector->ihevc_intra_pred_luma_mode_11_to_17_fptr = &ihevc_intra_pred_luma_mode_11_to_17_a9q;
ps_func_selector->ihevc_intra_pred_luma_mode_19_to_25_fptr = &ihevc_intra_pred_luma_mode_19_to_25_a9q;
ps_func_selector->ihevc_intra_pred_luma_dc_fptr = &ihevc_intra_pred_luma_dc_a9q;
ps_func_selector->ihevc_intra_pred_luma_horz_fptr = &ihevc_intra_pred_luma_horz_a9q;
ps_func_selector->ihevc_intra_pred_luma_mode2_fptr = &ihevc_intra_pred_luma_mode2_a9q;
ps_func_selector->ihevc_intra_pred_luma_mode_18_34_fptr = &ihevc_intra_pred_luma_mode_18_34_a9q;
ps_func_selector->ihevc_intra_pred_luma_mode_27_to_33_fptr = &ihevc_intra_pred_luma_mode_27_to_33_a9q;
ps_func_selector->ihevc_intra_pred_luma_mode_3_to_9_fptr = &ihevc_intra_pred_luma_mode_3_to_9_a9q;
ps_func_selector->ihevc_intra_pred_luma_planar_fptr = &ihevc_intra_pred_luma_planar_a9q;
ps_func_selector->ihevc_intra_pred_luma_ver_fptr = &ihevc_intra_pred_luma_ver_a9q;
ps_func_selector->ihevc_itrans_4x4_ttype1_fptr = &ihevc_itrans_4x4_ttype1;
ps_func_selector->ihevc_itrans_4x4_fptr = &ihevc_itrans_4x4;
ps_func_selector->ihevc_itrans_8x8_fptr = &ihevc_itrans_8x8;
ps_func_selector->ihevc_itrans_16x16_fptr = &ihevc_itrans_16x16;
ps_func_selector->ihevc_itrans_32x32_fptr = &ihevc_itrans_32x32;
ps_func_selector->ihevc_itrans_res_4x4_ttype1_fptr = &ihevc_itrans_res_4x4_ttype1;
ps_func_selector->ihevc_itrans_res_4x4_fptr = &ihevc_itrans_res_4x4;
ps_func_selector->ihevc_itrans_res_8x8_fptr = &ihevc_itrans_res_8x8;
ps_func_selector->ihevc_itrans_res_16x16_fptr = &ihevc_itrans_res_16x16;
ps_func_selector->ihevc_itrans_res_32x32_fptr = &ihevc_itrans_res_32x32;
ps_func_selector->ihevc_itrans_recon_4x4_ttype1_fptr = &ihevc_itrans_recon_4x4_ttype1_a9q;
ps_func_selector->ihevc_itrans_recon_4x4_fptr = &ihevc_itrans_recon_4x4_a9q;
ps_func_selector->ihevc_itrans_recon_8x8_fptr = &ihevc_itrans_recon_8x8_a9q;
ps_func_selector->ihevc_itrans_recon_16x16_fptr = &ihevc_itrans_recon_16x16_a9q;
ps_func_selector->ihevc_itrans_recon_32x32_fptr = &ihevc_itrans_recon_32x32_a9q;
ps_func_selector->ihevc_chroma_itrans_recon_4x4_fptr = &ihevc_chroma_itrans_recon_4x4;
ps_func_selector->ihevc_chroma_itrans_recon_8x8_fptr = &ihevc_chroma_itrans_recon_8x8;
ps_func_selector->ihevc_chroma_itrans_recon_16x16_fptr = &ihevc_chroma_itrans_recon_16x16;
ps_func_selector->ihevc_chroma_itrans_recon_32x32_fptr = &ihevc_chroma_itrans_recon_32x32;
ps_func_selector->ihevc_recon_4x4_ttype1_fptr = &ihevc_recon_4x4_ttype1;
ps_func_selector->ihevc_recon_4x4_fptr = &ihevc_recon_4x4;
ps_func_selector->ihevc_recon_8x8_fptr = &ihevc_recon_8x8;
ps_func_selector->ihevc_recon_16x16_fptr = &ihevc_recon_16x16;
ps_func_selector->ihevc_recon_32x32_fptr = &ihevc_recon_32x32;
ps_func_selector->ihevc_chroma_recon_4x4_fptr = &ihevc_chroma_recon_4x4;
ps_func_selector->ihevc_chroma_recon_8x8_fptr = &ihevc_chroma_recon_8x8;
ps_func_selector->ihevc_chroma_recon_16x16_fptr = &ihevc_chroma_recon_16x16;
ps_func_selector->ihevc_chroma_recon_32x32_fptr = &ihevc_chroma_recon_32x32;
ps_func_selector->ihevc_memcpy_mul_8_fptr = &ihevc_memcpy_mul_8_a9q;
ps_func_selector->ihevc_memcpy_fptr = &ihevc_memcpy_a9q;
ps_func_selector->ihevc_memset_mul_8_fptr = &ihevc_memset_mul_8_a9q;
ps_func_selector->ihevc_memset_fptr = &ihevc_memset_a9q;
ps_func_selector->ihevc_memset_16bit_mul_8_fptr = &ihevc_memset_16bit_mul_8_a9q;
ps_func_selector->ihevc_memset_16bit_fptr = &ihevc_memset_16bit_a9q;
ps_func_selector->ihevc_pad_left_luma_fptr = &ihevc_pad_left_luma_a9q;
ps_func_selector->ihevc_pad_left_chroma_fptr = &ihevc_pad_left_chroma_a9q;
ps_func_selector->ihevc_pad_right_luma_fptr = &ihevc_pad_right_luma_a9q;
ps_func_selector->ihevc_pad_right_chroma_fptr = &ihevc_pad_right_chroma_a9q;
ps_func_selector->ihevc_weighted_pred_bi_fptr = &ihevc_weighted_pred_bi_a9q;
ps_func_selector->ihevc_weighted_pred_bi_default_fptr = &ihevc_weighted_pred_bi_default_a9q;
ps_func_selector->ihevc_weighted_pred_uni_fptr = &ihevc_weighted_pred_uni_a9q;
ps_func_selector->ihevc_weighted_pred_chroma_bi_fptr = &ihevc_weighted_pred_chroma_bi_neonintr;
ps_func_selector->ihevc_weighted_pred_chroma_bi_default_fptr = &ihevc_weighted_pred_chroma_bi_default_neonintr;
ps_func_selector->ihevc_weighted_pred_chroma_uni_fptr = &ihevc_weighted_pred_chroma_uni_neonintr;
ps_func_selector->ihevc_sao_band_offset_luma_fptr = &ihevc_sao_band_offset_luma_a9q;
ps_func_selector->ihevc_sao_band_offset_chroma_fptr = &ihevc_sao_band_offset_chroma_a9q;
ps_func_selector->ihevc_sao_edge_offset_class0_fptr = &ihevc_sao_edge_offset_class0_a9q;
ps_func_selector->ihevc_sao_edge_offset_class0_chroma_fptr = &ihevc_sao_edge_offset_class0_chroma_a9q;
ps_func_selector->ihevc_sao_edge_offset_class1_fptr = &ihevc_sao_edge_offset_class1_a9q;
ps_func_selector->ihevc_sao_edge_offset_class1_chroma_fptr = &ihevc_sao_edge_offset_class1_chroma_a9q;
ps_func_selector->ihevc_sao_edge_offset_class2_fptr = &ihevc_sao_edge_offset_class2_a9q;
ps_func_selector->ihevc_sao_edge_offset_class2_chroma_fptr = &ihevc_sao_edge_offset_class2_chroma_a9q;
ps_func_selector->ihevc_sao_edge_offset_class3_fptr = &ihevc_sao_edge_offset_class3_a9q;
ps_func_selector->ihevc_sao_edge_offset_class3_chroma_fptr = &ihevc_sao_edge_offset_class3_chroma_a9q;
ps_func_selector->ihevcd_fmt_conv_420sp_to_420sp_fptr = &ihevcd_fmt_conv_420sp_to_420sp;
ps_func_selector->ihevcd_fmt_conv_420sp_to_420p_fptr = &ihevcd_fmt_conv_420sp_to_420p_a9q;
ps_func_selector->ihevcd_fmt_conv_444sp_to_444p_fptr = &ihevcd_fmt_conv_444sp_to_444p;
ps_func_selector->ihevcd_itrans_recon_dc_luma_fptr = &ihevcd_itrans_recon_dc_luma_a9q;
ps_func_selector->ihevcd_itrans_recon_dc_chroma_fptr = &ihevcd_itrans_recon_dc_chroma_a9q;
ps_func_selector->ihevcd_itrans_res_dc_fptr = &ihevcd_itrans_res_dc;
ps_codec->s_func_selector.ihevc_deblk_chroma_horz_fptr = &ihevc_deblk_chroma_horz_a9q;
ps_codec->s_func_selector.ihevc_deblk_chroma_vert_fptr = &ihevc_deblk_chroma_vert_a9q;
ps_codec->s_func_selector.ihevc_deblk_luma_vert_fptr = &ihevc_deblk_luma_vert_a9q;
ps_codec->s_func_selector.ihevc_deblk_luma_horz_fptr = &ihevc_deblk_luma_horz_a9q;
ps_codec->s_func_selector.ihevc_inter_pred_chroma_copy_fptr = &ihevc_inter_pred_chroma_copy_a9q;
ps_codec->s_func_selector.ihevc_inter_pred_chroma_copy_w16out_fptr = &ihevc_inter_pred_chroma_copy_w16out_a9q;
ps_codec->s_func_selector.ihevc_inter_pred_chroma_horz_fptr = &ihevc_inter_pred_chroma_horz_a9q;
ps_codec->s_func_selector.ihevc_inter_pred_chroma_horz_w16out_fptr = &ihevc_inter_pred_chroma_horz_w16out_a9q;
ps_codec->s_func_selector.ihevc_inter_pred_chroma_vert_fptr = &ihevc_inter_pred_chroma_vert_a9q;
ps_codec->s_func_selector.ihevc_inter_pred_chroma_vert_w16inp_fptr = &ihevc_inter_pred_chroma_vert_w16inp_a9q;
ps_codec->s_func_selector.ihevc_inter_pred_chroma_vert_w16inp_w16out_fptr = &ihevc_inter_pred_chroma_vert_w16inp_w16out_a9q;
ps_codec->s_func_selector.ihevc_inter_pred_chroma_vert_w16out_fptr = &ihevc_inter_pred_chroma_vert_w16out_a9q;
ps_codec->s_func_selector.ihevc_inter_pred_luma_horz_fptr = &ihevc_inter_pred_luma_horz_a9q;
ps_codec->s_func_selector.ihevc_inter_pred_luma_vert_fptr = &ihevc_inter_pred_luma_vert_a9q;
ps_codec->s_func_selector.ihevc_inter_pred_luma_vert_w16out_fptr = &ihevc_inter_pred_luma_vert_w16out_a9q;
ps_codec->s_func_selector.ihevc_inter_pred_luma_vert_w16inp_fptr = &ihevc_inter_pred_luma_vert_w16inp_a9q;
ps_codec->s_func_selector.ihevc_inter_pred_luma_copy_fptr = &ihevc_inter_pred_luma_copy_a9q;
ps_codec->s_func_selector.ihevc_inter_pred_luma_copy_w16out_fptr = &ihevc_inter_pred_luma_copy_w16out_a9q;
ps_codec->s_func_selector.ihevc_inter_pred_luma_horz_w16out_fptr = &ihevc_inter_pred_luma_horz_w16out_a9q;
ps_codec->s_func_selector.ihevc_inter_pred_luma_vert_w16inp_w16out_fptr = &ihevc_inter_pred_luma_vert_w16inp_w16out_a9q;
ps_codec->s_func_selector.ihevc_intra_pred_chroma_ref_substitution_fptr = &ihevc_intra_pred_chroma_ref_substitution_a9q;
ps_codec->s_func_selector.ihevc_intra_pred_luma_ref_substitution_fptr = &ihevc_intra_pred_luma_ref_substitution_a9q;
ps_codec->s_func_selector.ihevc_intra_pred_luma_ref_subst_all_avlble_fptr = &ihevc_intra_pred_luma_ref_subst_all_avlble;
ps_codec->s_func_selector.ihevc_intra_pred_ref_filtering_fptr = &ihevc_intra_pred_ref_filtering_neonintr;
ps_codec->s_func_selector.ihevc_intra_pred_chroma_dc_fptr = &ihevc_intra_pred_chroma_dc_a9q;
ps_codec->s_func_selector.ihevc_intra_pred_chroma_horz_fptr = &ihevc_intra_pred_chroma_horz_a9q;
ps_codec->s_func_selector.ihevc_intra_pred_chroma_mode2_fptr = &ihevc_intra_pred_chroma_mode2_a9q;
ps_codec->s_func_selector.ihevc_intra_pred_chroma_mode_18_34_fptr = &ihevc_intra_pred_chroma_mode_18_34_a9q;
ps_codec->s_func_selector.ihevc_intra_pred_chroma_mode_27_to_33_fptr = &ihevc_intra_pred_chroma_mode_27_to_33_a9q;
ps_codec->s_func_selector.ihevc_intra_pred_chroma_mode_3_to_9_fptr = &ihevc_intra_pred_chroma_mode_3_to_9_a9q;
ps_codec->s_func_selector.ihevc_intra_pred_chroma_planar_fptr = &ihevc_intra_pred_chroma_planar_a9q;
ps_codec->s_func_selector.ihevc_intra_pred_chroma_ver_fptr = &ihevc_intra_pred_chroma_ver_a9q;
ps_codec->s_func_selector.ihevc_intra_pred_chroma_mode_11_to_17_fptr = &ihevc_intra_pred_chroma_mode_11_to_17_a9q;
ps_codec->s_func_selector.ihevc_intra_pred_chroma_mode_19_to_25_fptr = &ihevc_intra_pred_chroma_mode_19_to_25_a9q;
ps_codec->s_func_selector.ihevc_intra_pred_luma_mode_11_to_17_fptr = &ihevc_intra_pred_luma_mode_11_to_17_a9q;
ps_codec->s_func_selector.ihevc_intra_pred_luma_mode_19_to_25_fptr = &ihevc_intra_pred_luma_mode_19_to_25_a9q;
ps_codec->s_func_selector.ihevc_intra_pred_luma_dc_fptr = &ihevc_intra_pred_luma_dc_a9q;
ps_codec->s_func_selector.ihevc_intra_pred_luma_horz_fptr = &ihevc_intra_pred_luma_horz_a9q;
ps_codec->s_func_selector.ihevc_intra_pred_luma_mode2_fptr = &ihevc_intra_pred_luma_mode2_a9q;
ps_codec->s_func_selector.ihevc_intra_pred_luma_mode_18_34_fptr = &ihevc_intra_pred_luma_mode_18_34_a9q;
ps_codec->s_func_selector.ihevc_intra_pred_luma_mode_27_to_33_fptr = &ihevc_intra_pred_luma_mode_27_to_33_a9q;
ps_codec->s_func_selector.ihevc_intra_pred_luma_mode_3_to_9_fptr = &ihevc_intra_pred_luma_mode_3_to_9_a9q;
ps_codec->s_func_selector.ihevc_intra_pred_luma_planar_fptr = &ihevc_intra_pred_luma_planar_a9q;
ps_codec->s_func_selector.ihevc_intra_pred_luma_ver_fptr = &ihevc_intra_pred_luma_ver_a9q;
ps_codec->s_func_selector.ihevc_itrans_4x4_ttype1_fptr = &ihevc_itrans_4x4_ttype1;
ps_codec->s_func_selector.ihevc_itrans_4x4_fptr = &ihevc_itrans_4x4;
ps_codec->s_func_selector.ihevc_itrans_8x8_fptr = &ihevc_itrans_8x8;
ps_codec->s_func_selector.ihevc_itrans_16x16_fptr = &ihevc_itrans_16x16;
ps_codec->s_func_selector.ihevc_itrans_32x32_fptr = &ihevc_itrans_32x32;
ps_codec->s_func_selector.ihevc_itrans_recon_4x4_ttype1_fptr = &ihevc_itrans_recon_4x4_ttype1_a9q;
ps_codec->s_func_selector.ihevc_itrans_recon_4x4_fptr = &ihevc_itrans_recon_4x4_a9q;
ps_codec->s_func_selector.ihevc_itrans_recon_8x8_fptr = &ihevc_itrans_recon_8x8_a9q;
ps_codec->s_func_selector.ihevc_itrans_recon_16x16_fptr = &ihevc_itrans_recon_16x16_a9q;
ps_codec->s_func_selector.ihevc_itrans_recon_32x32_fptr = &ihevc_itrans_recon_32x32_a9q;
ps_codec->s_func_selector.ihevc_chroma_itrans_recon_4x4_fptr = &ihevc_chroma_itrans_recon_4x4;
ps_codec->s_func_selector.ihevc_chroma_itrans_recon_8x8_fptr = &ihevc_chroma_itrans_recon_8x8;
ps_codec->s_func_selector.ihevc_chroma_itrans_recon_16x16_fptr = &ihevc_chroma_itrans_recon_16x16;
ps_codec->s_func_selector.ihevc_recon_4x4_ttype1_fptr = &ihevc_recon_4x4_ttype1;
ps_codec->s_func_selector.ihevc_recon_4x4_fptr = &ihevc_recon_4x4;
ps_codec->s_func_selector.ihevc_recon_8x8_fptr = &ihevc_recon_8x8;
ps_codec->s_func_selector.ihevc_recon_16x16_fptr = &ihevc_recon_16x16;
ps_codec->s_func_selector.ihevc_recon_32x32_fptr = &ihevc_recon_32x32;
ps_codec->s_func_selector.ihevc_chroma_recon_4x4_fptr = &ihevc_chroma_recon_4x4;
ps_codec->s_func_selector.ihevc_chroma_recon_8x8_fptr = &ihevc_chroma_recon_8x8;
ps_codec->s_func_selector.ihevc_chroma_recon_16x16_fptr = &ihevc_chroma_recon_16x16;
ps_codec->s_func_selector.ihevc_memcpy_mul_8_fptr = &ihevc_memcpy_mul_8_a9q;
ps_codec->s_func_selector.ihevc_memcpy_fptr = &ihevc_memcpy_a9q;
ps_codec->s_func_selector.ihevc_memset_mul_8_fptr = &ihevc_memset_mul_8_a9q;
ps_codec->s_func_selector.ihevc_memset_fptr = &ihevc_memset_a9q;
ps_codec->s_func_selector.ihevc_memset_16bit_mul_8_fptr = &ihevc_memset_16bit_mul_8_a9q;
ps_codec->s_func_selector.ihevc_memset_16bit_fptr = &ihevc_memset_16bit_a9q;
ps_codec->s_func_selector.ihevc_pad_left_luma_fptr = &ihevc_pad_left_luma_a9q;
ps_codec->s_func_selector.ihevc_pad_left_chroma_fptr = &ihevc_pad_left_chroma_a9q;
ps_codec->s_func_selector.ihevc_pad_right_luma_fptr = &ihevc_pad_right_luma_a9q;
ps_codec->s_func_selector.ihevc_pad_right_chroma_fptr = &ihevc_pad_right_chroma_a9q;
ps_codec->s_func_selector.ihevc_weighted_pred_bi_fptr = &ihevc_weighted_pred_bi_a9q;
ps_codec->s_func_selector.ihevc_weighted_pred_bi_default_fptr = &ihevc_weighted_pred_bi_default_a9q;
ps_codec->s_func_selector.ihevc_weighted_pred_uni_fptr = &ihevc_weighted_pred_uni_a9q;
ps_codec->s_func_selector.ihevc_weighted_pred_chroma_bi_fptr = &ihevc_weighted_pred_chroma_bi_neonintr;
ps_codec->s_func_selector.ihevc_weighted_pred_chroma_bi_default_fptr = &ihevc_weighted_pred_chroma_bi_default_neonintr;
ps_codec->s_func_selector.ihevc_weighted_pred_chroma_uni_fptr = &ihevc_weighted_pred_chroma_uni_neonintr;
ps_codec->s_func_selector.ihevc_sao_band_offset_luma_fptr = &ihevc_sao_band_offset_luma_a9q;
ps_codec->s_func_selector.ihevc_sao_band_offset_chroma_fptr = &ihevc_sao_band_offset_chroma_a9q;
ps_codec->s_func_selector.ihevc_sao_edge_offset_class0_fptr = &ihevc_sao_edge_offset_class0_a9q;
ps_codec->s_func_selector.ihevc_sao_edge_offset_class0_chroma_fptr = &ihevc_sao_edge_offset_class0_chroma_a9q;
ps_codec->s_func_selector.ihevc_sao_edge_offset_class1_fptr = &ihevc_sao_edge_offset_class1_a9q;
ps_codec->s_func_selector.ihevc_sao_edge_offset_class1_chroma_fptr = &ihevc_sao_edge_offset_class1_chroma_a9q;
ps_codec->s_func_selector.ihevc_sao_edge_offset_class2_fptr = &ihevc_sao_edge_offset_class2_a9q;
ps_codec->s_func_selector.ihevc_sao_edge_offset_class2_chroma_fptr = &ihevc_sao_edge_offset_class2_chroma_a9q;
ps_codec->s_func_selector.ihevc_sao_edge_offset_class3_fptr = &ihevc_sao_edge_offset_class3_a9q;
ps_codec->s_func_selector.ihevc_sao_edge_offset_class3_chroma_fptr = &ihevc_sao_edge_offset_class3_chroma_a9q;
ps_codec->s_func_selector.ihevcd_fmt_conv_420sp_to_rgba8888_fptr = &ihevcd_fmt_conv_420sp_to_rgba8888_a9q;
ps_codec->s_func_selector.ihevcd_fmt_conv_420sp_to_rgb565_fptr = &ihevcd_fmt_conv_420sp_to_rgb565;
ps_codec->s_func_selector.ihevcd_fmt_conv_420sp_to_420sp_fptr = &ihevcd_fmt_conv_420sp_to_420sp;
ps_codec->s_func_selector.ihevcd_fmt_conv_420sp_to_420p_fptr = &ihevcd_fmt_conv_420sp_to_420p_a9q;
ps_codec->s_func_selector.ihevcd_itrans_recon_dc_luma_fptr = &ihevcd_itrans_recon_dc_luma_a9q;
ps_codec->s_func_selector.ihevcd_itrans_recon_dc_chroma_fptr = &ihevcd_itrans_recon_dc_chroma_a9q;
}

Some files were not shown because too many files have changed in this diff Show more