Merge pull request #3 from rockchip-linux/develop

update

CHANGELOG.md

@@ -1,3 +1,272 @@
+## 1.0.7 (2024-09-04)
+### Feature
+- [rc_smt]: Add rc container for smart mode
+- [vepu580]: Optimization to improve VMAF
+- [vepu580]: Optimize hal processing for smart encoding
+- [vepu580]: Add qpmap and rc container interface
+- [vepu510]: Add anti-smear regs setup for H.264
+- [vepu510]: Add H.264 tuning setup
+- [vepu510]: Sync code from enc_tune branch
+- [vepu510]: Sync code from enc_tune branch
+- [vepu510]: Sync code from enc_tune branch
+- [mpp_trie]: Add trie context filling feature
+- [mpp_trie]: Add trie tag and shrink feature
+- [h264d]: support hdr meta parse
+- [h265e]: Support force mark & use ltr
+- [vpu_api]: support yuv444sp decode ouput pixel format
+
+### Fix
+- [h265d]: fix infochange loss when two sps continuous
+- [hal_h264e]: Fix CAVLC encode smartP stream err
+- [mpi_enc_test]: Remove redundant code about smart encoding
+- [h264e_sps]: fix the default value of max mv length
+- [enc_roi]: Fix cu_map init in vepu_54x_roi
+- [hal_vp9]: Optimize prob memory usage
+- [hal_h265d]: Allow reference missing for GDR
+- [osal]: Fix mpp_mem single instance issue
+- [hal_vp9d_com]: Fixed memory leak issue
+- [hal_h265d]: Avoid risk of segment fault
+- [hal_h265d]: fix error slot index marking
+- [h265d]: Adjust condition of scan type judgement
+- [mpp_hdr]: Fix buffer overflow issue
+- [mpp_buffer]: Synchronous log addition point
+- [hal_vepu]: fix split regs assignment
+- [vepu580]: poll max set to 1 on split out lowdelay mode
+- [mpp_common]: fix compile err on F_DUPFD_CLOEXEC not defined
+- [h265d]: return error on sps/pps read failure
+- [build]: The first toolchains is selected by default
+- [265e]:Fix the st refernce frame err in tsvc
+- [av1d]: when MetaData found then it is new frame
+- [m2vd]: Fix seq_head check error
+- [h265e_vepu510]: Fix a memory leak
+- [h265d]: auto output frame in dpb when ready
+- [m2vd]: Remove ref frame when info changed
+- [mpp_meta]: Missing data in the instance
+- [mpp_bitread]: Fix negative shift error
+- [osal]: fix 128 odd plus 64 bytes alignment
+- [h265d_parser]: Fix fmt configuration issue
+- [hal_av1d_vdpu383]: modify av1 segid wr/rd base config
+- [h265d_parser]: Fix fmt configuration issue
+- [hal_av1d_vdpu383]: add segid reg base config
+
+### Docs
+- Update 1.0.7 CHANGELOG.md
+- [readme]: Add more repo info
+
+### Refactor
+- [mpp_cfg]: Refactor MppTrie and string cfg
+
+### Chore
+- [mpp_mem]: Add mpp_realloc_size
+- [mpp_cfg]: Remove some unused code
+- fix compile warning
+
+## 1.0.6 (2024-06-12)
+### Feature
+- [vdpu383]: refine rcb info setup
+- [enc_265]: Support get Largest Code Unit size
+- [mpp_dec_cfg]: Add disable dpb check config
+- [vdpu383]: support 8K downscale mode
+
+### Fix
+- [drm]: Fix permission check issue on GKI kernel
+- [hal_h265e]: Amend 510 tid and sync cache
+- [hal_h265e]: Fix nalu type avoid stream warning
+- [h265e]: Fix vps/sps max temparal layers val
+- [hal_jpeg_vdpu1]: fix dec failed on RK3036 problem
+- [osal]: rv1109/rv1126 vcodec_type mismatch problem
+- [h264e_vepu2]: Adjust inter favor table
+- [h264d]: fix drop packets after reset when err stream
+- [h265d]: Allow filtering of consecutive start code
+- [hal_h264d_vdpu383]: fix spspps update issue
+- [mpp]: fix mpp frame leak when async enc
+- [enc]: Add use_lt_idx to output packet meta
+- [hal_h265e]: fix sse_sum get err
+- [mpp_enc_async]: fix mpp packet leak when thread quit
+- [enc_roi]: Support ROI cfg under CQP mode
+- [hal]: Fix the lib interdependence issue
+- [vepu_510]: fix same log type when enc feedback
+- [mpp_buffer]: fix dec/inc ref_count in multi threads
+- [mpp_enc_async]: fix debreath not work on async flow
+- [base]: fix AV1 and AVS2 string info missing problem
+- [mpp]: Add encoder input/output pts log
+- [hal_vepu580/510]: fix split out err when pass1 frame
+- [hal]: Fix target link issue
+- [hal_enc]: Fix lib dependency issue
+- [hal_h265d_vdpu383]: fix ref_err mark for special poc
+- [rc2_test]: fix pkt buffer overflow error
+- [enc_utils]: Support read odd resolution image
+- [allocator]: fix on invalid DMA heap allocator
+- [hal_h265e_vepu580]: fix reg config err for 2pass
+- [jpegd_vdpu]: Adjust file dump path
+- [mpp_common]: fix 128 odd plus 64 alignment
+- [cmake]: fix static build
+- [vdpu383]: Update vdpu383 error detection
+
+### Docs
+- Update 1.0.6 CHANGELOG.md
+
+### Refactor
+- [hal_jpegd]: init devices at hal_jpegd_api
+- [dec]: get deocder capability via common routine
+- [hal_av1d]: Migrate av1d from vpu to rkdec
+
+### Chore
+- [h265d]: Reduce malloc/free frequency of vps
+- [mpp_service]: fix typo err
+- [hal_h265d]: use INT_MAX for poc distance initiation
+- [cmake]: remove duplicate code
+
+## 1.0.5 (2024-04-19)
+### Feature
+- [vdpu383]: align hor stride to 128 odds + 64 byte
+- [vdpu383]: support 2x2 scale down
+- [mpp_buffer]: Add MppBuffer attach/detach func
+- [mpp_dev]: Add fd attach/detach operation
+- [vdpp]: Add libvdpp for hwpq
+- [vdpp]: Add capacity check function
+- [cmake]: Add building static library
+- [vdpp_test]: Add vdpp slt testcase
+- [av1d]: Add tile4x4 frame format support
+- [mpp_enc_cfg]: Add H.265 tier config
+- [jpeg]: Add VPU720 JPEG support
+- [enc]: Add config entry for output chroma format
+- [vdpu383]: Add vdpu383 av1 module
+- [vdpu383]: Add vdpu383 vp9 module
+- [vdpu383]: Add vdpu383 avs2 module
+- [vdpu383]: Add vdpu383 H.264 module
+- [vdpu383]: Add vdpu383 H.265 module
+- [vdpu383]: Add vdpu383 common module
+- [vdpp]: Add vdpp2 for rk3576
+- [vdpp]: Add vdpp module and vdpp_test
+- [vepu_510]: Add vepu510 h265e support
+- [vepu_510]: Add vepu510 h264e support
+- [mpp_frame]: Add tile format flag
+- [vepu_510] add vepu 510 common for H264 & h265
+- [mpp_soc]: support rk3576 soc
+
+### Fix
+- [avs2d_vdpu383]: Optimise dec result
+- [vdpu383]: Fix compiler warning
+- [vdpp]: Fix dmsr reg size imcompat error
+- [vdpu383]: hor stride alignment fix for vdpu383
+- [h265d_ref]: fix set fbc output fmt not effect issue
+- [vdpu383]: Fix memory out of bounds issue
+- [h264d_vdpu383]: Fix global parameter config issue
+- [avs2_parse]: add colorspace config to mpp_frame
+- [hal_h264e]:fix crash after init vepu buffer failure
+- [vpu_api]: Fix frame format and eos cfg
+- [av1d_vdpu383]: fix fbc hor_stride error
+- [av1d_parser]: fix fmt error for 10bit HDR source
+- [avs2d]: fix stuck when seq_end shows at pkt tail
+- [av1d_vdpu]: Fix forced 8bit output failure issue
+- [enc_async]: Invalidate cache for output buffer
+- [hal_av1d_vdpu383]: memleak for cdf_bufs
+- [av1d_vdpu383]: fix rcb buffer size
+- [vp9d_vdpu383]: Fix segid config issue
+- [vepu510]: Add split low delay output mode support
+- [avs2d_vdpu383]: Fix declaring shadow local variables issue
+- [av1]: Fix global config issue
+- [hal_av1d]: Delte cdf unused value
+- [av1]: Fix av1 display abnormality issue
+- [avs2d]: Remove a unnecessary log
+- [vepu510]: Adjust regs assignment
+- [hal_jpegd]: Add stream buffer flush
+- [265e_api]: Support cons_intra_pred_flag cfg
+- [mpp_enc]: Add device attach/detach on enc flow
+- [mpp_dec]: Add device attach/detach on dec flow
+- [vdpp]: Add error detection
+- [hal_265e_510]: modify srgn_max & rime_lvl val
+- [vdpu383]: spspps data not need copy all range ppsid
+- [vpu_api_legacy]: fix frame CodingType err
+- [av1]: Fix 10bit source display issue
+- [mpp_enc]: Expand the hdr_buf size
+- [av1]: Fix delta_q value read issue
+- [vdpu383]: Enable error detection
+- [ext_dma]: fix mmap permission error
+- [jpege_vpu720]: sync cache before return task
+- [mpp_buffer]: fix buffer type assigning
+- [vepu510]: Configure reg of Subjective param
+- [vepu510]: Checkout and optimize 510 reg.h
+- [mpp_dec]: Optimize HDR meta process
+- [av1d]: Fix scanlist calc issue
+- [h265e]: fix the profile tier cfg
+- [av1d]: Fix av1d ref stride error
+- [hal_h265e_vepu510]: Add cudecis reg cfg
+- [av1d]: Only rk3588 support 10bit translate to 8bit
+- [vp9d]: Fix vp9 hor stride issue
+- [rc]: Add i quality delta cfg on fixqp mode
+- [hal_h265d]: Fix filter col rcb buffer size calc
+- [av1d]: Fix compiler warning
+- [h264d]: Fix error mvc stream crash issue
+- [hal_h264e]: Fix qp err when fixqp mode
+- [h264d]: Fix H.264 error chroma_format_idc
+- [vdpu383]: Fix av1 rkfbc output error
+- [av1d]: Fix compatibility issues
+- [hal_h264d_vdpu383]: Reduce mpp_put_bits calls
+- Fix clerical error
+- [avs2d]: Fix get ref_frm slot idx error
+- [vdpu383]: Fix av1 global params bit pos issue
+- [vdpp]: fix sharp config error
+- [hal_av1d]: fix av1 dec err for rk3588
+- [vdpu383]: Fix av1 global params issue
+- [vepu510]: Fix camera record stuck issue
+- [utils]: fix read and write some YUV format
+- [mpp_bitput]: fix put bits overflow
+- [mpp_service]: fix rcb info env config
+- [vepu510]: Fix compile warning
+- [hal_vp9d]: fix colmv size calculator err
+- [avsd]: Fix the ref_frm slot idx erro in fast mode.
+
+### Docs
+- Update 1.0.5 CHANGELOG.md
+- [mpp_frame]: Add MppFrameFormat description
+
+### Refactor
+- [hal_av2sd]: refactor hal_api assign flow
+- [hal_h264d]: refactor hal_api assign flow
+- Using soc_type for compare instead of soc_name
+
+### Chore
+- [hal_h264e]: clean some unused code
+
+## 1.0.4 (2024-02-07)
+### Feature
+- [vpu_api_legacy]: Support RGB24 setup
+- [avsd]: keep codec type if not avs+
+- [mpi_enc_test]: add YUV400 fmt support
+- [mpp_enc]: Add YUV400 support for vepu580/540
+
+### Fix
+- [h265e]: fix hw stream size check error
+- [hal_vdpu]: unify colmv buffer size calculation
+- [vproc]: Fix deadlock in vproc thread
+- [h265e]: disable tmvp by default
+- [h265e]: Amend temporal_id to stream
+- [mpp_dump]: add YUV420SP_10BIT format dump
+- [hal_h265d]: Fix register length for rk3328/rk3328H
+- [hal_avsd]: Fix crash on no buffer decoded
+- [mpp_enc]: allow vp8 to cfg force idr frame
+- [m2vd]: fix unindentical of input and output pts list
+- [h265e_vepu580]: fix SIGSEGV when reencoding
+- [mpp_dmabuf]: fix align cache line size calculate err
+- [h265e_vepu580]: flush cache for the first tile
+- [dmabuf]: Disable dmabuf partial sync function
+- [iep_test]: use internal buffer group
+- [common]: Add mpp_dup function
+- [h265e]: Adapter RK3528 when encoding P frame skip
+- [h265e]: fix missing end_of_slice_segment_flag problem
+- [hal_av1d_vdpu]: change rkv_hor_align to 16 align
+- [av1d_parser]: set color info per frame
+- [jpegd]: add sof marker check when parser done
+
+### Docs
+- Update 1.0.4 CHANGELOG.md
+
+### Chore
+- [script]: add rebuild and clean for build
+- [mpp_enc_roi_utils]: change file format dos to unix
+
 ## 1.0.3 (2023-12-08)
 ### Feature
 - [dec_test]: Add buffer mode option

CMakeLists.txt

@@ -171,6 +171,8 @@ if(GCC)
     add_definitions(-fPIC)
     # disable exception for C++
     set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fno-exceptions -fno-rtti")
+    # save intermediate files
+    # add_definitions(-save-temps)
 
     # for libary linking
     set(BEGIN_WHOLE_ARCHIVE -Wl,--whole-archive)
@@ -292,8 +294,7 @@ endif(MSVC)
 # ----------------------------------------------------------------------------
 # Share library option
 # ----------------------------------------------------------------------------
-option(ENABLE_STATIC "Build shared library" ON)
-option(ENABLE_SHARED "Build shared library" OFF)
+option(BUILD_SHARED_LIBS "Build shared library" ON)
 
 # ----------------------------------------------------------------------------
 # scan all LOG_TAG for log information and generate module header file

build/android/env_setup.sh

@@ -170,15 +170,16 @@ case ${NDK_COUNT} in
         echo "use ndk: ${ANDROID_NDK}"
         ;;
     *)
-        read -p "select [1-${NDK_COUNT}] ndk used for compiling: " -ra NDK_INTPUT
+        read -p "select [1-${NDK_COUNT}] ndk used for compiling, def[1]: " -ra NDK_INTPUT
+        NDK_INTPUT=${NDK_INTPUT:-1}
 
         NDK_INDEX=0
 
         for NDK_PATH in ${NDK_OPTION[@]};
         do
             NDK_INDEX=$[${NDK_INDEX}+1]
 
-            if [ "${NDK_INDEX}" -eq "${NDK_INTPUT}" ]; then
+            if [ "${NDK_INDEX}" == "${NDK_INTPUT}" ]; then
                 echo "${NDK_INTPUT} - ${NDK_PATH} selected as ANDROID_NDK"
                 ANDROID_NDK=${NDK_PATH}
                 FOUND_NDK=1

debian/compat

@@ -1 +1 @@
-9
+10

debian/control

@@ -1,7 +1,7 @@
 Source: mpp
 Priority: optional
 Maintainer: Randy Li <[email protected]>
-Build-Depends: debhelper (>= 9), cmake
+Build-Depends: debhelper (>= 10), cmake, dh-exec
 Standards-Version: 3.9.5
 Section: libs
 Homepage: http://www.rock-chips.com

doc/Rockchip_Developer_Guide_MPP_CN.md

@@ -613,12 +613,12 @@ MPP_RET mpp_enc_cfg_get_st(MppEncCfg cfg, const char *name, void *val);
 | rc:bps_target           | S32  | RK_S32                              | 表示CBR模式下的目标码率。                                    |
 | rc:bps_max              | S32  | RK_S32                              | 表示VBR/AVBR模式下的最高码率。                               |
 | rc:bps_min              | S32  | RK_S32                              | 表示VBR/AVBR模式下的最低码率。                               |
-| rc:fps_in_flex          | S32  | RK_S32                              | 表示输入帧率是否可变的标志位，默认为0。 为0表示输入帧率固定，帧率计算方式为： fps_in_num / fps_in_denorm，可以表示分数帧率。 为1表示输入帧率可变。可变帧率的情况下，帧率不固定，对应的码率计算与分配的规则变为按实际时间进行计算。 |
+| rc:fps_in_flex          | S32  | RK_S32                              | 表示输入帧率是否可变的标志位，默认为0。 为0表示输入帧率固定，帧率计算方式为： fps_in_num / fps_in_denom，可以表示分数帧率。 为1表示输入帧率可变。可变帧率的情况下，帧率不固定，对应的码率计算与分配的规则变为按实际时间进行计算。 |
 | rc:fps_in_num           | S32  | RK_S32                              | 表示输入帧率分数值的分子部分，默认值为30。                   |
-| rc:fps_in_denorm        | S32  | RK_S32                              | 表示输入帧率分数值的分母部分，默认值为1。                    |
-| rc:fps_out_flex         | S32  | RK_S32                              | 表示输出帧率是否可变的标志位，默认为0。 为0表示输出帧率固定，帧率计算方式为： fps_out_num / fps_out_denorm，可以表示分数帧率。 为1表示输出帧率可变。可变帧率的情况下，帧率不固定，对应的码流输出时间按实际时间进行计算。 |
+| rc:fps_in_denom         | S32  | RK_S32                              | 表示输入帧率分数值的分母部分，默认值为1。                    |
+| rc:fps_out_flex         | S32  | RK_S32                              | 表示输出帧率是否可变的标志位，默认为0。 为0表示输出帧率固定，帧率计算方式为： fps_out_num / fps_out_denom，可以表示分数帧率。 为1表示输出帧率可变。可变帧率的情况下，帧率不固定，对应的码流输出时间按实际时间进行计算。 |
 | rc:fps_out_num          | S32  | RK_S32                              | 表示输出帧率分数值的分子部分，默认值为30。                   |
-| rc:fps_out_denorm       | S32  | RK_S32                              | 表示输出帧率分数值的分母部分，默认值为1。                    |
+| rc:fps_out_denom        | S32  | RK_S32                              | 表示输出帧率分数值的分母部分，默认值为1。                    |
 | rc:gop                  | S32  | RK_S32                              | 表示Group Of Picture，即两个I帧之间的间隔，含义如下： 0 – 表示只有一个I帧，其他帧均为P帧。 1 – 表示全为I帧。 2 – 表示序列为I P I P I P… 3 – 表示序列为I P P I P P I P P… 一般情况下，gop应配置为输出帧率的整数倍，默认值为两倍输出帧率。 |
 | rc:max_reenc_times      | U32  | RK_U32                              | 表示一帧图像最大重编码次数，默认值为1。在低延时输出模式下，max_reenc_times只能配置为0。 |
 | rc:priority             | U32  | MppEncRcPriority                    | 表示超大帧重编优先级。 0 – 表示目标码率优先。 1 – 表示超大帧阈值优先。 此优先级只在超大帧重编时有效。 ![](media/Rockchip_Developer_Guide_MPP/MPP_MppEncRcPriority.png) |