视频编解码初步

字节　半字　字　双字　四字x86: b w d q dqmips: b h w d q源码目录：run-openh264/: 运行程序的脚本文件openh264/: 我修改的openh264-nomodify/: 未修改，直接用git clone http://cgit.loongnix.org/cgit/openh264 下载的。25的服务器上：scp -r run-openh264/ loongson@10.20.42.30:~cd /home/build/scp -r toolchain/ root@10.20.42.30:/opt/本地机器上：cd /opt/mv toolchain/* .tar -xf gdb.tar.gztar -xf gcc-4.8.3-1.loongson.3.tar.bz2mv gcc-4.8.3-1.loongson.3 gcc-4.8.3vim bash，将其内容拷贝到文件/etc/bashrcsource /etc/bashrcwhich gcccd openh264-nomodify/make -j4：编译cp run-openh264/* ../g.sh：运行，生成norc_noaq_nodenois_noscene8.264md5sum norc_noaq_nodenois_noscene8.264 ：生成校验码ffplay norc_noaq_nodenois_noscene8.264mplayer norc_noaq_nodenois_noscene8.264提交：git format-patch -1patch -p1 < 0001-codec-processing-add-function-VAACalcSadSsd_mmi.patchgit push origin HEAD:refs/for/loongson-v1.5.3-Firefox39git pullgit reset --hard 版本号readelf -h h264decmake -B ENABLE64BIT=No BUILDTYPE=Release调试：gdb ./h264encr 参数x/10i $pcinfo reg浅谈C/C++堆栈指引： /article/2957868.html社区源码库：　http://gerrit.loongnix.org/Gerrit系统使用帮助： http://10.2.5.21/showdoc/index.php/6修改函数：参考x86汇编：codec/processing/src/x86/vaa.asm中函数VAACalcSadBgd_sse2编写mips汇编：codec/processing/src/vaacalc/vaacalculation.cpp中函数VAACalcSadBgd_mmi对应的C源代码：codec/processing/src/vaacalc/vaacalcfuncs.cpp中函数VAACalcSadBgd_c龙芯开源社区: http://www.loongnix.org/cgitMMX指令集: /article/9399399.htmlSSE2指令集: http://blog.sina.com.cn/s/blog_54d5cd260100oduv.htmlGCC在C语言中内嵌汇编 asm volatile : /article/1907472.htmlcodec/encoder/core/src/deblocking.cppcodec/common/x86/deblock.asmcodec/common/src/deblocking_common.cpp

long long f0=0,f2=0,f4=0,f6=0,f8=0,f10=0,f12=0,f14=0;
long long f16=0,f18=0,f20=0,f22=0,f24=0,f26=0,f28=0,f30=0;

"dmfc1     %[f0] , $f0                                  \n\t"
"dmfc1     %[f2] , $f2                                  \n\t"
"dmfc1     %[f4] , $f4                                  \n\t"
"dmfc1     %[f6] , $f6                                  \n\t"
"dmfc1     %[f8] , $f8                                  \n\t"
"dmfc1     %[f10] , $f10                                  \n\t"
"dmfc1     %[f12] , $f12                                  \n\t"
"dmfc1     %[f14] , $f14                                  \n\t"
"dmfc1     %[f16] , $f16                                  \n\t"
"dmfc1     %[f18] , $f18                                  \n\t"
"dmfc1     %[f20] , $f20                                  \n\t"
"dmfc1     %[f22] , $f22                                  \n\t"
"dmfc1     %[f24] , $f24                                  \n\t"
"dmfc1     %[f26] , $f26                                  \n\t"
"dmfc1     %[f28] , $f28                                  \n\t"
"dmfc1     %[f30] , $f30                                  \n\t"

:  [pPixCb]"+&r"((unsigned char *)pPixCb), [pPixCr]"+&r"((unsigned char *)pPixCr),
[f0]"=&r"(f0), [f2]"=&r"(f2), [f4]"=&r"(f4), [f6]"=&r"(f6),
[f8]"=&r"(f8), [f10]"=&r"(f10), [f12]"=&r"(f12), [f14]"=&r"(f14),
[f16]"=&r"(f16), [f18]"=&r"(f18), [f20]"=&r"(f20), [f22]"=&r"(f22),
[f24]"=&r"(f24), [f26]"=&r"(f26), [f28]"=&r"(f28), [f30]"=&r"(f30)
:  [iStride]"r"((int)iStride), [iAlpha]"r"(iAlpha),
[iBeta]"r"(iBeta), [pTC]"r"((unsigned char *)pTC), [tmp]"r"((unsigned char *)tmp)
: "memory", "$8", "$9", "$10", "$11", "$12", "$13", "$15", "$f0", "$f2", "$f4", "$f6", "$f8", "$f10", "$f12",
"$f14", "$f16", "$f18", "$f20", "$f22", "$f24", "$f26", "$f28", "$f30"
);

printf("f0:%x\tf2:%x\tf4:%x\tf6:%x\tf8:%x\tf10:%x\tf12:%x\tf14:%x\n",f0,f2,f4,f6,f8,f10,f12,f14);
printf("\n");
printf("f16:%x\tf18:%x\tf20:%x\tf22:%x\tf24:%x\tf26:%x\tf28:%x\tf30:%x\n",f16,f18,f20,f22,f24,f26,f28,f30);
printf("\n");

asm volatile (
".set     arch=loongson3a                             \n\t"