diff options
Diffstat (limited to 'libavutil')
287 files changed, 24530 insertions, 2545 deletions
diff --git a/libavutil/Makefile b/libavutil/Makefile index 11d91794cb..a35deb6fec 100644 --- a/libavutil/Makefile +++ b/libavutil/Makefile @@ -1,7 +1,10 @@ +include $(SUBDIR)../config.mak + NAME = avutil HEADERS = adler32.h \ aes.h \ + aes_ctr.h \ attributes.h \ audio_fifo.h \ avassert.h \ @@ -9,12 +12,16 @@ HEADERS = adler32.h \ avutil.h \ base64.h \ blowfish.h \ + bprint.h \ bswap.h \ buffer.h \ + cast5.h \ + camellia.h \ channel_layout.h \ common.h \ cpu.h \ crc.h \ + des.h \ dict.h \ display.h \ downmix_info.h \ @@ -23,6 +30,7 @@ HEADERS = adler32.h \ fifo.h \ file.h \ frame.h \ + hash.h \ hmac.h \ hwcontext.h \ hwcontext_cuda.h \ @@ -35,40 +43,60 @@ HEADERS = adler32.h \ log.h \ macros.h \ mathematics.h \ + mastering_display_metadata.h \ md5.h \ mem.h \ + motion_vector.h \ + murmur3.h \ opt.h \ parseutils.h \ pixdesc.h \ + pixelutils.h \ pixfmt.h \ random_seed.h \ + rc4.h \ rational.h \ replaygain.h \ + ripemd.h \ samplefmt.h \ sha.h \ + sha512.h \ stereo3d.h \ + threadmessage.h \ time.h \ + timecode.h \ + timestamp.h \ + tree.h \ + twofish.h \ version.h \ xtea.h \ + tea.h \ HEADERS-$(CONFIG_LZO) += lzo.h +HEADERS-$(CONFIG_OPENCL) += opencl.h + ARCH_HEADERS = bswap.h \ intmath.h \ intreadwrite.h \ timer.h \ -BUILT_HEADERS = avconfig.h +BUILT_HEADERS = avconfig.h \ + ffversion.h OBJS = adler32.o \ aes.o \ - atomic.o \ + aes_ctr.o \ audio_fifo.o \ avstring.o \ base64.o \ blowfish.o \ + bprint.o \ buffer.o \ + cast5.o \ + camellia.o \ channel_layout.o \ + color_utils.o \ cpu.o \ crc.o \ des.o \ @@ -81,45 +109,66 @@ OBJS = adler32.o \ file.o \ file_open.o \ float_dsp.o \ + fixed_dsp.o \ frame.o \ + hash.o \ hmac.o \ hwcontext.o \ imgutils.o \ + integer.o \ intmath.o \ lfg.o \ lls.o \ log.o \ log2_tab.o \ mathematics.o \ + mastering_display_metadata.o \ md5.o \ mem.o \ + murmur3.o \ opt.o \ parseutils.o \ pixdesc.o \ + pixelutils.o \ random_seed.o \ rational.o \ + reverse.o \ rc4.o \ + ripemd.o \ samplefmt.o \ sha.o \ + sha512.o \ stereo3d.o \ + threadmessage.o \ time.o \ + timecode.o \ tree.o \ + twofish.o \ utils.o \ + xga_font_data.o \ xtea.o \ + tea.o \ + +OBJS-$(!HAVE_ATOMICS_NATIVE) += atomic.o \ OBJS-$(CONFIG_CUDA) += hwcontext_cuda.o OBJS-$(CONFIG_LZO) += lzo.o +OBJS-$(CONFIG_OPENCL) += opencl.o opencl_internal.o OBJS-$(CONFIG_VAAPI) += hwcontext_vaapi.o OBJS-$(CONFIG_VDPAU) += hwcontext_vdpau.o OBJS += $(COMPAT_OBJS:%=../compat/%) +# Windows resource file +SLIBOBJS-$(HAVE_GNU_WINDRES) += avutilres.o + SKIPHEADERS-$(CONFIG_CUDA) += hwcontext_cuda.h SKIPHEADERS-$(CONFIG_VAAPI) += hwcontext_vaapi.h SKIPHEADERS-$(CONFIG_VDPAU) += hwcontext_vdpau.h SKIPHEADERS-$(HAVE_ATOMICS_GCC) += atomic_gcc.h SKIPHEADERS-$(HAVE_ATOMICS_SUNCC) += atomic_suncc.h SKIPHEADERS-$(HAVE_ATOMICS_WIN32) += atomic_win32.h +SKIPHEADERS-$(CONFIG_OPENCL) += opencl.h TESTPROGS = adler32 \ aes \ @@ -127,18 +176,49 @@ TESTPROGS = adler32 \ avstring \ base64 \ blowfish \ + bprint \ + cast5 \ + camellia \ + color_utils \ cpu \ crc \ des \ + dict \ + display \ + error \ eval \ + file \ fifo \ float_dsp \ + hash \ hmac \ lfg \ lls \ + log \ md5 \ + murmur3 \ opt \ + pca \ parseutils \ + pixdesc \ + pixelutils \ + random_seed \ + rational \ + ripemd \ sha \ + sha512 \ + softfloat \ tree \ + twofish \ + utf8 \ xtea \ + tea \ + +TESTPROGS-$(HAVE_LZO1X_999_COMPRESS) += lzo + +TOOLS = crypto_bench ffhash ffeval ffescape + +tools/crypto_bench$(EXESUF): ELIBS += $(if $(VERSUS),$(subst +, -l,+$(VERSUS)),) +tools/crypto_bench$(EXESUF): CFLAGS += -DUSE_EXT_LIBS=0$(if $(VERSUS),$(subst +,+USE_,+$(VERSUS)),) + +$(SUBDIR)lzo-test$(EXESUF): ELIBS = -llzo2 diff --git a/libavutil/aarch64/asm.S b/libavutil/aarch64/asm.S index 6a7f506d13..ff34e7a5e5 100644 --- a/libavutil/aarch64/asm.S +++ b/libavutil/aarch64/asm.S @@ -1,20 +1,20 @@ /* * Copyright (c) 2008 Mans Rullgard <mans@mansr.com> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/aarch64/bswap.h b/libavutil/aarch64/bswap.h index 0c001cec19..1e735c52cf 100644 --- a/libavutil/aarch64/bswap.h +++ b/libavutil/aarch64/bswap.h @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/aarch64/cpu.c b/libavutil/aarch64/cpu.c index 37a7d8defa..8ef077aaea 100644 --- a/libavutil/aarch64/cpu.c +++ b/libavutil/aarch64/cpu.c @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/aarch64/cpu.h b/libavutil/aarch64/cpu.h index f5b1d89132..cf1b9cc516 100644 --- a/libavutil/aarch64/cpu.h +++ b/libavutil/aarch64/cpu.h @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/aarch64/float_dsp_init.c b/libavutil/aarch64/float_dsp_init.c index 37d34c96d0..4325071821 100644 --- a/libavutil/aarch64/float_dsp_init.c +++ b/libavutil/aarch64/float_dsp_init.c @@ -2,20 +2,20 @@ * ARM NEON optimised Float DSP functions * Copyright (c) 2008 Mans Rullgard <mans@mansr.com> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/aarch64/float_dsp_neon.S b/libavutil/aarch64/float_dsp_neon.S index 776542c4a9..02d790c0cc 100644 --- a/libavutil/aarch64/float_dsp_neon.S +++ b/libavutil/aarch64/float_dsp_neon.S @@ -3,20 +3,20 @@ * Copyright (c) 2008 Mans Rullgard <mans@mansr.com> * Copyright (c) 2014 Janne Grunau <janne-libav@jannau.net> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/aarch64/neontest.h b/libavutil/aarch64/neontest.h index de841bbbae..2d0fc19994 100644 --- a/libavutil/aarch64/neontest.h +++ b/libavutil/aarch64/neontest.h @@ -3,23 +3,26 @@ * Copyright (c) 2008 Ramiro Polla <ramiro.polla@gmail.com> * Copyright (c) 2013 Martin Storsjo * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ +#ifndef AVUTIL_AARCH64_NEONTEST_H +#define AVUTIL_AARCH64_NEONTEST_H + #include <inttypes.h> #include <stdint.h> #include <stdlib.h> @@ -63,3 +66,5 @@ int __real_ ## func; \ int __wrap_ ## func; \ int __wrap_ ## func + +#endif /* AVUTIL_AARCH64_NEONTEST_H */ diff --git a/libavutil/aarch64/timer.h b/libavutil/aarch64/timer.h index 382cfd959c..b570039416 100644 --- a/libavutil/aarch64/timer.h +++ b/libavutil/aarch64/timer.h @@ -1,20 +1,20 @@ /* * Copyright (c) 2015 Janne Grunau <janne-libav@jannau.net> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/adler32-test.c b/libavutil/adler32-test.c index ab109f9175..f93cf9d449 100644 --- a/libavutil/adler32-test.c +++ b/libavutil/adler32-test.c @@ -1,21 +1,22 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ +// LCOV_EXCL_START #include <string.h> #include "log.h" @@ -29,7 +30,7 @@ static volatile int checksum; int main(int argc, char **argv) { int i; - char data[LEN]; + uint8_t data[LEN]; av_log_set_level(AV_LOG_DEBUG); @@ -49,3 +50,4 @@ int main(int argc, char **argv) av_log(NULL, AV_LOG_DEBUG, "%X (expected 50E6E508)\n", checksum); return checksum == 0x50e6e508 ? 0 : 1; } +// LCOV_EXCL_STOP diff --git a/libavutil/adler32.c b/libavutil/adler32.c index 8dfe2ce451..c87d5e261c 100644 --- a/libavutil/adler32.c +++ b/libavutil/adler32.c @@ -32,6 +32,8 @@ #include "config.h" #include "adler32.h" +#include "common.h" +#include "intreadwrite.h" #define BASE 65521L /* largest prime smaller than 65536 */ @@ -46,16 +48,46 @@ unsigned long av_adler32_update(unsigned long adler, const uint8_t * buf, unsigned long s2 = adler >> 16; while (len > 0) { -#if CONFIG_SMALL +#if HAVE_FAST_64BIT && HAVE_FAST_UNALIGNED && !CONFIG_SMALL + unsigned len2 = FFMIN((len-1) & ~7, 23*8); + if (len2) { + uint64_t a1= 0; + uint64_t a2= 0; + uint64_t b1= 0; + uint64_t b2= 0; + len -= len2; + s2 += s1*len2; + while (len2 >= 8) { + uint64_t v = AV_RN64(buf); + a2 += a1; + b2 += b1; + a1 += v &0x00FF00FF00FF00FF; + b1 += (v>>8)&0x00FF00FF00FF00FF; + len2 -= 8; + buf+=8; + } + + //We combine the 8 interleaved adler32 checksums without overflows + //Decreasing the number of iterations would allow below code to be + //simplified but would likely be slower due to the fewer iterations + //of the inner loop + s1 += ((a1+b1)*0x1000100010001)>>48; + s2 += ((((a2&0xFFFF0000FFFF)+(b2&0xFFFF0000FFFF)+((a2>>16)&0xFFFF0000FFFF)+((b2>>16)&0xFFFF0000FFFF))*0x800000008)>>32) +#if HAVE_BIGENDIAN + + 2*((b1*0x1000200030004)>>48) + + ((a1*0x1000100010001)>>48) + + 2*((a1*0x0000100020003)>>48); +#else + + 2*((a1*0x4000300020001)>>48) + + ((b1*0x1000100010001)>>48) + + 2*((b1*0x3000200010000)>>48); +#endif + } +#else while (len > 4 && s2 < (1U << 31)) { DO4(buf); len -= 4; } -#else - while (len > 16 && s2 < (1U << 31)) { - DO16(buf); - len -= 16; - } #endif DO1(buf); len--; s1 %= BASE; diff --git a/libavutil/adler32.h b/libavutil/adler32.h index 53e3dbf51d..0dc69ec0a8 100644 --- a/libavutil/adler32.h +++ b/libavutil/adler32.h @@ -1,20 +1,20 @@ /* * copyright (c) 2006 Mans Rullgard * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/aes-test.c b/libavutil/aes-test.c index 92404e4cc2..a71b7fee28 100644 --- a/libavutil/aes-test.c +++ b/libavutil/aes-test.c @@ -1,23 +1,24 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ #include "aes.c" +// LCOV_EXCL_START #include <string.h> #include "lfg.h" @@ -32,7 +33,7 @@ int main(int argc, char **argv) { 0x10, 0xa5, 0x88, 0x69, 0xd7, 0x4b, 0xe5, 0xa3, 0x74, 0xcf, 0x86, 0x7c, 0xfb, 0x47, 0x38, 0x59 } }; - uint8_t pt[16], rpt[2][16] = { + uint8_t pt[32], rpt[2][16] = { { 0x6a, 0x84, 0x86, 0x7c, 0xd7, 0x7e, 0x12, 0xad, 0x07, 0xea, 0x1b, 0xe8, 0x95, 0xc5, 0x3f, 0xa3 }, { 0 } @@ -43,7 +44,8 @@ int main(int argc, char **argv) { 0x6d, 0x25, 0x1e, 0x69, 0x44, 0xb0, 0x51, 0xe0, 0x4e, 0xaa, 0x6f, 0xb4, 0xdb, 0xf7, 0x84, 0x65 } }; - uint8_t temp[16]; + uint8_t temp[32]; + uint8_t iv[2][16]; int err = 0; av_log_set_level(AV_LOG_DEBUG); @@ -64,20 +66,27 @@ int main(int argc, char **argv) AVAES ae, ad; AVLFG prng; - av_aes_init(&ae, "PI=3.141592654..", 128, 0); - av_aes_init(&ad, "PI=3.141592654..", 128, 1); + av_aes_init(&ae, (const uint8_t*)"PI=3.141592654..", 128, 0); + av_aes_init(&ad, (const uint8_t*)"PI=3.141592654..", 128, 1); av_lfg_init(&prng, 1); for (i = 0; i < 10000; i++) { - for (j = 0; j < 16; j++) + for (j = 0; j < 32; j++) pt[j] = av_lfg_get(&prng); + for (j = 0; j < 16; j++) + iv[0][j] = iv[1][j] = av_lfg_get(&prng); { START_TIMER; - av_aes_crypt(&ae, temp, pt, 1, NULL, 0); + av_aes_crypt(&ae, temp, pt, 2, iv[0], 0); + if (!(i & (i - 1))) + av_log(NULL, AV_LOG_ERROR, "%02X %02X %02X %02X\n", + temp[0], temp[5], temp[10], temp[15]); + av_aes_crypt(&ad, temp, temp, 2, iv[1], 1); + av_aes_crypt(&ae, temp, pt, 2, NULL, 0); if (!(i & (i - 1))) av_log(NULL, AV_LOG_ERROR, "%02X %02X %02X %02X\n", temp[0], temp[5], temp[10], temp[15]); - av_aes_crypt(&ad, temp, temp, 1, NULL, 1); + av_aes_crypt(&ad, temp, temp, 2, NULL, 1); STOP_TIMER("aes"); } for (j = 0; j < 16; j++) { @@ -90,3 +99,4 @@ int main(int argc, char **argv) } return err; } +// LCOV_EXCL_STOP diff --git a/libavutil/aes.c b/libavutil/aes.c index e93fb0565c..397ea77389 100644 --- a/libavutil/aes.c +++ b/libavutil/aes.c @@ -3,42 +3,30 @@ * * some optimization ideas from aes128.c by Reimar Doeffinger * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ #include "common.h" +#include "aes.h" +#include "aes_internal.h" #include "intreadwrite.h" #include "timer.h" -#include "aes.h" -typedef union { - uint64_t u64[2]; - uint32_t u32[4]; - uint8_t u8x4[4][4]; - uint8_t u8[16]; -} av_aes_block; - -typedef struct AVAES { - // Note: round_key[16] is accessed in the init code, but this only - // overwrites state, which does not matter (see also commit ba554c0). - av_aes_block round_key[15]; - av_aes_block state[2]; - int rounds; -} AVAES; +const int av_aes_size= sizeof(AVAES); struct AVAES *av_aes_alloc(void) { @@ -60,9 +48,9 @@ static uint32_t dec_multbl[4][256]; #endif #if HAVE_BIGENDIAN -# define ROT(x, s) ((x >> s) | (x << (32-s))) +# define ROT(x, s) (((x) >> (s)) | ((x) << (32-(s)))) #else -# define ROT(x, s) ((x << s) | (x >> (32-s))) +# define ROT(x, s) (((x) << (s)) | ((x) >> (32-(s)))) #endif static inline void addkey(av_aes_block *dst, const av_aes_block *src, @@ -127,7 +115,7 @@ static inline void mix(av_aes_block state[2], uint32_t multbl[][256], int s1, in state[0].u32[3] = mix_core(multbl, src[3][0], src[s1 - 1][1], src[1][2], src[s3 - 1][3]); } -static inline void crypt(AVAES *a, int s, const uint8_t *sbox, +static inline void aes_crypt(AVAES *a, int s, const uint8_t *sbox, uint32_t multbl[][256]) { int r; @@ -140,31 +128,44 @@ static inline void crypt(AVAES *a, int s, const uint8_t *sbox, subshift(&a->state[0], s, sbox); } -void av_aes_crypt(AVAES *a, uint8_t *dst, const uint8_t *src, - int count, uint8_t *iv, int decrypt) +static void aes_encrypt(AVAES *a, uint8_t *dst, const uint8_t *src, + int count, uint8_t *iv, int rounds) +{ + while (count--) { + addkey_s(&a->state[1], src, &a->round_key[rounds]); + if (iv) + addkey_s(&a->state[1], iv, &a->state[1]); + aes_crypt(a, 2, sbox, enc_multbl); + addkey_d(dst, &a->state[0], &a->round_key[0]); + if (iv) + memcpy(iv, dst, 16); + src += 16; + dst += 16; + } +} + +static void aes_decrypt(AVAES *a, uint8_t *dst, const uint8_t *src, + int count, uint8_t *iv, int rounds) { while (count--) { - addkey_s(&a->state[1], src, &a->round_key[a->rounds]); - if (decrypt) { - crypt(a, 0, inv_sbox, dec_multbl); - if (iv) { - addkey_s(&a->state[0], iv, &a->state[0]); - memcpy(iv, src, 16); - } - addkey_d(dst, &a->state[0], &a->round_key[0]); - } else { - if (iv) - addkey_s(&a->state[1], iv, &a->state[1]); - crypt(a, 2, sbox, enc_multbl); - addkey_d(dst, &a->state[0], &a->round_key[0]); - if (iv) - memcpy(iv, dst, 16); + addkey_s(&a->state[1], src, &a->round_key[rounds]); + aes_crypt(a, 0, inv_sbox, dec_multbl); + if (iv) { + addkey_s(&a->state[0], iv, &a->state[0]); + memcpy(iv, src, 16); } + addkey_d(dst, &a->state[0], &a->round_key[0]); src += 16; dst += 16; } } +void av_aes_crypt(AVAES *a, uint8_t *dst, const uint8_t *src, + int count, uint8_t *iv, int decrypt) +{ + a->crypt(a, dst, src, count, iv, a->rounds); +} + static void init_multbl2(uint32_t tbl[][256], const int c[4], const uint8_t *log8, const uint8_t *alog8, const uint8_t *sbox) @@ -200,6 +201,8 @@ int av_aes_init(AVAES *a, const uint8_t *key, int key_bits, int decrypt) uint8_t log8[256]; uint8_t alog8[512]; + a->crypt = decrypt ? aes_decrypt : aes_encrypt; + if (!enc_multbl[FF_ARRAY_ELEMS(enc_multbl) - 1][FF_ARRAY_ELEMS(enc_multbl[0]) - 1]) { j = 1; for (i = 0; i < 255; i++) { @@ -223,7 +226,7 @@ int av_aes_init(AVAES *a, const uint8_t *key, int key_bits, int decrypt) } if (key_bits != 128 && key_bits != 192 && key_bits != 256) - return -1; + return AVERROR(EINVAL); a->rounds = rounds; @@ -262,3 +265,4 @@ int av_aes_init(AVAES *a, const uint8_t *key, int key_bits, int decrypt) return 0; } + diff --git a/libavutil/aes.h b/libavutil/aes.h index 5b451241e4..09efbda107 100644 --- a/libavutil/aes.h +++ b/libavutil/aes.h @@ -1,20 +1,20 @@ /* * copyright (c) 2007 Michael Niedermayer <michaelni@gmx.at> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -32,6 +32,8 @@ * @{ */ +extern const int av_aes_size; + struct AVAES; /** diff --git a/libavutil/aes_ctr.c b/libavutil/aes_ctr.c new file mode 100644 index 0000000000..e9c568fe0d --- /dev/null +++ b/libavutil/aes_ctr.c @@ -0,0 +1,129 @@ +/* + * AES-CTR cipher + * Copyright (c) 2015 Eran Kornblau <erankor at gmail dot com> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "common.h" +#include "aes_ctr.h" +#include "aes.h" +#include "random_seed.h" + +#define AES_BLOCK_SIZE (16) + +typedef struct AVAESCTR { + struct AVAES* aes; + uint8_t counter[AES_BLOCK_SIZE]; + uint8_t encrypted_counter[AES_BLOCK_SIZE]; + int block_offset; +} AVAESCTR; + +struct AVAESCTR *av_aes_ctr_alloc(void) +{ + return av_mallocz(sizeof(struct AVAESCTR)); +} + +void av_aes_ctr_set_iv(struct AVAESCTR *a, const uint8_t* iv) +{ + memcpy(a->counter, iv, AES_CTR_IV_SIZE); + memset(a->counter + AES_CTR_IV_SIZE, 0, sizeof(a->counter) - AES_CTR_IV_SIZE); + a->block_offset = 0; +} + +const uint8_t* av_aes_ctr_get_iv(struct AVAESCTR *a) +{ + return a->counter; +} + +void av_aes_ctr_set_random_iv(struct AVAESCTR *a) +{ + uint32_t iv[2]; + + iv[0] = av_get_random_seed(); + iv[1] = av_get_random_seed(); + + av_aes_ctr_set_iv(a, (uint8_t*)iv); +} + +int av_aes_ctr_init(struct AVAESCTR *a, const uint8_t *key) +{ + a->aes = av_aes_alloc(); + if (!a->aes) { + return AVERROR(ENOMEM); + } + + av_aes_init(a->aes, key, 128, 0); + + memset(a->counter, 0, sizeof(a->counter)); + a->block_offset = 0; + + return 0; +} + +void av_aes_ctr_free(struct AVAESCTR *a) +{ + if (a) { + av_freep(&a->aes); + av_free(a); + } +} + +static void av_aes_ctr_increment_be64(uint8_t* counter) +{ + uint8_t* cur_pos; + + for (cur_pos = counter + 7; cur_pos >= counter; cur_pos--) { + (*cur_pos)++; + if (*cur_pos != 0) { + break; + } + } +} + +void av_aes_ctr_increment_iv(struct AVAESCTR *a) +{ + av_aes_ctr_increment_be64(a->counter); + memset(a->counter + AES_CTR_IV_SIZE, 0, sizeof(a->counter) - AES_CTR_IV_SIZE); + a->block_offset = 0; +} + +void av_aes_ctr_crypt(struct AVAESCTR *a, uint8_t *dst, const uint8_t *src, int count) +{ + const uint8_t* src_end = src + count; + const uint8_t* cur_end_pos; + uint8_t* encrypted_counter_pos; + + while (src < src_end) { + if (a->block_offset == 0) { + av_aes_crypt(a->aes, a->encrypted_counter, a->counter, 1, NULL, 0); + + av_aes_ctr_increment_be64(a->counter + 8); + } + + encrypted_counter_pos = a->encrypted_counter + a->block_offset; + cur_end_pos = src + AES_BLOCK_SIZE - a->block_offset; + cur_end_pos = FFMIN(cur_end_pos, src_end); + + a->block_offset += cur_end_pos - src; + a->block_offset &= (AES_BLOCK_SIZE - 1); + + while (src < cur_end_pos) { + *dst++ = *src++ ^ *encrypted_counter_pos++; + } + } +} diff --git a/libavutil/aes_ctr.h b/libavutil/aes_ctr.h new file mode 100644 index 0000000000..f596fa6a46 --- /dev/null +++ b/libavutil/aes_ctr.h @@ -0,0 +1,83 @@ +/* + * AES-CTR cipher + * Copyright (c) 2015 Eran Kornblau <erankor at gmail dot com> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#ifndef AVUTIL_AES_CTR_H +#define AVUTIL_AES_CTR_H + +#include <stdint.h> + +#include "attributes.h" +#include "version.h" + +#define AES_CTR_KEY_SIZE (16) +#define AES_CTR_IV_SIZE (8) + +struct AVAESCTR; + +/** + * Allocate an AVAESCTR context. + */ +struct AVAESCTR *av_aes_ctr_alloc(void); + +/** + * Initialize an AVAESCTR context. + * @param key encryption key, must have a length of AES_CTR_KEY_SIZE + */ +int av_aes_ctr_init(struct AVAESCTR *a, const uint8_t *key); + +/** + * Release an AVAESCTR context. + */ +void av_aes_ctr_free(struct AVAESCTR *a); + +/** + * Process a buffer using a previously initialized context. + * @param dst destination array, can be equal to src + * @param src source array, can be equal to dst + * @param size the size of src and dst + */ +void av_aes_ctr_crypt(struct AVAESCTR *a, uint8_t *dst, const uint8_t *src, int size); + +/** + * Get the current iv + */ +const uint8_t* av_aes_ctr_get_iv(struct AVAESCTR *a); + +/** + * Generate a random iv + */ +void av_aes_ctr_set_random_iv(struct AVAESCTR *a); + +/** + * Forcefully change the iv + */ +void av_aes_ctr_set_iv(struct AVAESCTR *a, const uint8_t* iv); + +/** + * Increment the top 64 bit of the iv (performed after each frame) + */ +void av_aes_ctr_increment_iv(struct AVAESCTR *a); + +/** + * @} + */ + +#endif /* AVUTIL_AES_CTR_H */ diff --git a/libavutil/aes_internal.h b/libavutil/aes_internal.h new file mode 100644 index 0000000000..494425878d --- /dev/null +++ b/libavutil/aes_internal.h @@ -0,0 +1,43 @@ +/* + * copyright (c) 2015 Rodger Combs <rodger.combs@gmail.com> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#ifndef AVUTIL_AES_INTERNAL_H +#define AVUTIL_AES_INTERNAL_H + +#include "mem.h" +#include <stdint.h> + +typedef union { + uint64_t u64[2]; + uint32_t u32[4]; + uint8_t u8x4[4][4]; + uint8_t u8[16]; +} av_aes_block; + +typedef struct AVAES { + // Note: round_key[16] is accessed in the init code, but this only + // overwrites state, which does not matter (see also commit ba554c0). + DECLARE_ALIGNED(16, av_aes_block, round_key)[15]; + DECLARE_ALIGNED(16, av_aes_block, state)[2]; + int rounds; + void (*crypt)(struct AVAES *a, uint8_t *dst, const uint8_t *src, int count, uint8_t *iv, int rounds); +} AVAES; + +#endif /* AVUTIL_AES_INTERNAL_H */ diff --git a/libavutil/arm/asm.S b/libavutil/arm/asm.S index 943c1cea3f..e9b0bcabdb 100644 --- a/libavutil/arm/asm.S +++ b/libavutil/arm/asm.S @@ -1,20 +1,20 @@ /* * Copyright (c) 2008 Mans Rullgard <mans@mansr.com> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/arm/bswap.h b/libavutil/arm/bswap.h index 8bc7d9a8f7..611ff0ad5b 100644 --- a/libavutil/arm/bswap.h +++ b/libavutil/arm/bswap.h @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -44,7 +44,7 @@ static av_always_inline av_const unsigned av_bswap16(unsigned x) } #endif -#if !AV_GCC_VERSION_AT_LEAST(4,5) +#if AV_GCC_VERSION_AT_MOST(4,4) #define av_bswap32 av_bswap32 static av_always_inline av_const uint32_t av_bswap32(uint32_t x) { @@ -60,7 +60,7 @@ static av_always_inline av_const uint32_t av_bswap32(uint32_t x) #endif /* HAVE_ARMV6_INLINE */ return x; } -#endif /* !AV_GCC_VERSION_AT_LEAST(4,5) */ +#endif /* AV_GCC_VERSION_AT_MOST(4,4) */ #endif /* __ARMCC_VERSION */ diff --git a/libavutil/arm/cpu.c b/libavutil/arm/cpu.c index 2effb72610..3889ef011c 100644 --- a/libavutil/arm/cpu.c +++ b/libavutil/arm/cpu.c @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -128,6 +128,12 @@ int ff_get_cpu_flags_arm(void) trickle down. */ if (flags & (AV_CPU_FLAG_VFPV3 | AV_CPU_FLAG_NEON)) flags |= AV_CPU_FLAG_ARMV6T2; + else if (flags & (AV_CPU_FLAG_ARMV6T2 | AV_CPU_FLAG_ARMV6)) + /* Some functions use the 'setend' instruction which is deprecated on ARMv8 + * and serializing on some ARMv7 cores. This ensures such functions + * are only enabled on ARMv6. */ + flags |= AV_CPU_FLAG_SETEND; + if (flags & AV_CPU_FLAG_ARMV6T2) flags |= AV_CPU_FLAG_ARMV6; @@ -147,7 +153,8 @@ int ff_get_cpu_flags_arm(void) AV_CPU_FLAG_ARMV6T2 * HAVE_ARMV6T2 | AV_CPU_FLAG_VFP * HAVE_VFP | AV_CPU_FLAG_VFPV3 * HAVE_VFPV3 | - AV_CPU_FLAG_NEON * HAVE_NEON; + AV_CPU_FLAG_NEON * HAVE_NEON | + AV_CPU_FLAG_SETEND * !(HAVE_NEON | HAVE_VFPV3); } #endif diff --git a/libavutil/arm/cpu.h b/libavutil/arm/cpu.h index 127993e5dd..eb64ed5fcd 100644 --- a/libavutil/arm/cpu.h +++ b/libavutil/arm/cpu.h @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -29,16 +29,11 @@ #define have_vfp(flags) CPUEXT(flags, VFP) #define have_vfpv3(flags) CPUEXT(flags, VFPV3) #define have_neon(flags) CPUEXT(flags, NEON) +#define have_setend(flags) CPUEXT(flags, SETEND) /* some functions use the VFPv2 vector mode which is deprecated in ARMv7-A * and might trap on such CPU depending on the OS configuration */ #define have_vfp_vm(flags) \ (HAVE_VFP && ((flags) & AV_CPU_FLAG_VFP_VM)) -/* Some functions use the 'setend' instruction which is deprecated on ARMv8 - * and serializing on some ARMv7 cores. This macro ensures such functions - * are only enabled on ARMv6. */ -#define have_setend(flags) \ - (have_armv6(flags) && !(have_vfpv3(flags) || have_neon(flags))) - #endif /* AVUTIL_ARM_CPU_H */ diff --git a/libavutil/arm/float_dsp_arm.h b/libavutil/arm/float_dsp_arm.h index ec925ecf5c..fe311cc0d2 100644 --- a/libavutil/arm/float_dsp_arm.h +++ b/libavutil/arm/float_dsp_arm.h @@ -1,20 +1,20 @@ /* * Copyright (c) 2009 Mans Rullgard <mans@mansr.com> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/arm/float_dsp_init_arm.c b/libavutil/arm/float_dsp_init_arm.c index ff0105c5c9..678762862e 100644 --- a/libavutil/arm/float_dsp_init_arm.c +++ b/libavutil/arm/float_dsp_init_arm.c @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/arm/float_dsp_init_neon.c b/libavutil/arm/float_dsp_init_neon.c index 617bf5d576..689aa77c7b 100644 --- a/libavutil/arm/float_dsp_init_neon.c +++ b/libavutil/arm/float_dsp_init_neon.c @@ -2,20 +2,20 @@ * ARM NEON optimised Float DSP functions * Copyright (c) 2008 Mans Rullgard <mans@mansr.com> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/arm/float_dsp_init_vfp.c b/libavutil/arm/float_dsp_init_vfp.c index 61ff2ed38e..e15abf3f54 100644 --- a/libavutil/arm/float_dsp_init_vfp.c +++ b/libavutil/arm/float_dsp_init_vfp.c @@ -1,20 +1,20 @@ /* * Copyright (c) 2008 Siarhei Siamashka <ssvb@users.sourceforge.net> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -32,7 +32,7 @@ void ff_vector_fmul_window_vfp(float *dst, const float *src0, void ff_vector_fmul_reverse_vfp(float *dst, const float *src0, const float *src1, int len); -void ff_butterflies_float_vfp(float *restrict v1, float *restrict v2, int len); +void ff_butterflies_float_vfp(float *av_restrict v1, float *av_restrict v2, int len); av_cold void ff_float_dsp_init_vfp(AVFloatDSPContext *fdsp, int cpu_flags) { diff --git a/libavutil/arm/float_dsp_neon.S b/libavutil/arm/float_dsp_neon.S index 559b565628..3823227608 100644 --- a/libavutil/arm/float_dsp_neon.S +++ b/libavutil/arm/float_dsp_neon.S @@ -2,20 +2,20 @@ * ARM NEON optimised Float DSP functions * Copyright (c) 2008 Mans Rullgard <mans@mansr.com> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/arm/float_dsp_vfp.S b/libavutil/arm/float_dsp_vfp.S index 9f920aae70..7db2452284 100644 --- a/libavutil/arm/float_dsp_vfp.S +++ b/libavutil/arm/float_dsp_vfp.S @@ -1,20 +1,20 @@ /* * Copyright (c) 2008 Siarhei Siamashka <ssvb@users.sourceforge.net> * - * This file is part of Libav. + * This file is part of FFmpeg * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/arm/intmath.h b/libavutil/arm/intmath.h index 2b15ba0593..65e42c5733 100644 --- a/libavutil/arm/intmath.h +++ b/libavutil/arm/intmath.h @@ -1,20 +1,20 @@ /* * Copyright (c) 2010 Mans Rullgard <mans@mansr.com> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -31,9 +31,9 @@ #if HAVE_ARMV6_INLINE #define av_clip_uint8 av_clip_uint8_arm -static av_always_inline av_const unsigned av_clip_uint8_arm(int a) +static av_always_inline av_const int av_clip_uint8_arm(int a) { - unsigned x; + int x; __asm__ ("usat %0, #8, %1" : "=r"(x) : "r"(a)); return x; } @@ -47,9 +47,9 @@ static av_always_inline av_const int av_clip_int8_arm(int a) } #define av_clip_uint16 av_clip_uint16_arm -static av_always_inline av_const unsigned av_clip_uint16_arm(int a) +static av_always_inline av_const int av_clip_uint16_arm(int a) { - unsigned x; + int x; __asm__ ("usat %0, #16, %1" : "=r"(x) : "r"(a)); return x; } diff --git a/libavutil/arm/intreadwrite.h b/libavutil/arm/intreadwrite.h index 6eff7332fb..60fc860cbb 100644 --- a/libavutil/arm/intreadwrite.h +++ b/libavutil/arm/intreadwrite.h @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -23,14 +23,14 @@ #include "config.h" #include "libavutil/attributes.h" -#if HAVE_FAST_UNALIGNED && HAVE_INLINE_ASM && !AV_GCC_VERSION_AT_LEAST(4,7) +#if HAVE_FAST_UNALIGNED && HAVE_INLINE_ASM && AV_GCC_VERSION_AT_MOST(4,6) #define AV_RN16 AV_RN16 static av_always_inline unsigned AV_RN16(const void *p) { const uint8_t *q = p; unsigned v; -#if !AV_GCC_VERSION_AT_LEAST(4,6) +#if AV_GCC_VERSION_AT_MOST(4,5) __asm__ ("ldrh %0, %1" : "=r"(v) : "m"(*(const uint16_t *)q)); #elif defined __thumb__ __asm__ ("ldrh %0, %1" : "=r"(v) : "m"(q[0]), "m"(q[1])); diff --git a/libavutil/arm/neontest.h b/libavutil/arm/neontest.h index eb46df8e0b..d75ab8380b 100644 --- a/libavutil/arm/neontest.h +++ b/libavutil/arm/neontest.h @@ -3,23 +3,26 @@ * Copyright (c) 2008 Ramiro Polla <ramiro.polla@gmail.com> * Copyright (c) 2013 Martin Storsjo * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ +#ifndef AVUTIL_ARM_NEONTEST_H +#define AVUTIL_ARM_NEONTEST_H + #include <inttypes.h> #include <stdint.h> #include <stdlib.h> @@ -60,3 +63,5 @@ int __real_ ## func; \ int __wrap_ ## func; \ int __wrap_ ## func + +#endif /* AVUTIL_ARM_NEONTEST_H */ diff --git a/libavutil/arm/timer.h b/libavutil/arm/timer.h index 4bca877d00..5e8bc8edd0 100644 --- a/libavutil/arm/timer.h +++ b/libavutil/arm/timer.h @@ -1,20 +1,20 @@ /* * Copyright (c) 2009 Mans Rullgard <mans@mansr.com> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/atomic-test.c b/libavutil/atomic-test.c index c0a89a461e..397d8c9285 100644 --- a/libavutil/atomic-test.c +++ b/libavutil/atomic-test.c @@ -1,24 +1,23 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ -#include <assert.h> - #include "atomic.h" +#include "avassert.h" int main(void) { @@ -26,10 +25,10 @@ int main(void) int res; res = avpriv_atomic_int_add_and_fetch(&val, 1); - assert(res == 2); + av_assert0(res == 2); avpriv_atomic_int_set(&val, 3); res = avpriv_atomic_int_get(&val); - assert(res == 3); + av_assert0(res == 3); return 0; } diff --git a/libavutil/atomic.c b/libavutil/atomic.c index e175a9f8ed..64cff2576f 100644 --- a/libavutil/atomic.c +++ b/libavutil/atomic.c @@ -1,20 +1,20 @@ /* * Copyright (c) 2012 Ronald S. Bultje <rsbultje@gmail.com> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -64,7 +64,7 @@ void *avpriv_atomic_ptr_cas(void * volatile *ptr, void *oldval, void *newval) void *ret; pthread_mutex_lock(&atomic_lock); ret = *ptr; - if (*ptr == oldval) + if (ret == oldval) *ptr = newval; pthread_mutex_unlock(&atomic_lock); return ret; diff --git a/libavutil/atomic.h b/libavutil/atomic.h index 57a0c19fcb..15906d24c9 100644 --- a/libavutil/atomic.h +++ b/libavutil/atomic.h @@ -1,20 +1,20 @@ /* * Copyright (c) 2012 Ronald S. Bultje <rsbultje@gmail.com> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/atomic_gcc.h b/libavutil/atomic_gcc.h index 9470e8e50b..5f9fc49ba0 100644 --- a/libavutil/atomic_gcc.h +++ b/libavutil/atomic_gcc.h @@ -1,20 +1,20 @@ /* * Copyright (c) 2012 Ronald S. Bultje <rsbultje@gmail.com> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -28,27 +28,40 @@ #define avpriv_atomic_int_get atomic_int_get_gcc static inline int atomic_int_get_gcc(volatile int *ptr) { +#if HAVE_ATOMIC_COMPARE_EXCHANGE + return __atomic_load_n(ptr, __ATOMIC_SEQ_CST); +#else __sync_synchronize(); return *ptr; +#endif } #define avpriv_atomic_int_set atomic_int_set_gcc static inline void atomic_int_set_gcc(volatile int *ptr, int val) { +#if HAVE_ATOMIC_COMPARE_EXCHANGE + __atomic_store_n(ptr, val, __ATOMIC_SEQ_CST); +#else *ptr = val; __sync_synchronize(); +#endif } #define avpriv_atomic_int_add_and_fetch atomic_int_add_and_fetch_gcc static inline int atomic_int_add_and_fetch_gcc(volatile int *ptr, int inc) { +#if HAVE_ATOMIC_COMPARE_EXCHANGE + return __atomic_add_fetch(ptr, inc, __ATOMIC_SEQ_CST); +#else return __sync_add_and_fetch(ptr, inc); +#endif } #define avpriv_atomic_ptr_cas atomic_ptr_cas_gcc static inline void *atomic_ptr_cas_gcc(void * volatile *ptr, void *oldval, void *newval) { +#if HAVE_SYNC_VAL_COMPARE_AND_SWAP #ifdef __ARMCC_VERSION // armcc will throw an error if ptr is not an integer type volatile uintptr_t *tmp = (volatile uintptr_t*)ptr; @@ -56,6 +69,10 @@ static inline void *atomic_ptr_cas_gcc(void * volatile *ptr, #else return __sync_val_compare_and_swap(ptr, oldval, newval); #endif +#else + __atomic_compare_exchange_n(ptr, &oldval, newval, 0, __ATOMIC_SEQ_CST, __ATOMIC_SEQ_CST); + return oldval; +#endif } #endif /* AVUTIL_ATOMIC_GCC_H */ diff --git a/libavutil/atomic_suncc.h b/libavutil/atomic_suncc.h index b78c37767a..a75a37b47e 100644 --- a/libavutil/atomic_suncc.h +++ b/libavutil/atomic_suncc.h @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/atomic_win32.h b/libavutil/atomic_win32.h index 00c66cfc1d..f7299336f6 100644 --- a/libavutil/atomic_win32.h +++ b/libavutil/atomic_win32.h @@ -1,26 +1,27 @@ /* * Copyright (c) 2012 Ronald S. Bultje <rsbultje@gmail.com> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ #ifndef AVUTIL_ATOMIC_WIN32_H #define AVUTIL_ATOMIC_WIN32_H +#define WIN32_LEAN_AND_MEAN #include <windows.h> #define avpriv_atomic_int_get atomic_int_get_win32 diff --git a/libavutil/attributes.h b/libavutil/attributes.h index d7f2bb5c6f..5c6b9deecb 100644 --- a/libavutil/attributes.h +++ b/libavutil/attributes.h @@ -1,20 +1,20 @@ /* * copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -27,11 +27,14 @@ #define AVUTIL_ATTRIBUTES_H #ifdef __GNUC__ -# define AV_GCC_VERSION_AT_LEAST(x,y) (__GNUC__ > x || __GNUC__ == x && __GNUC_MINOR__ >= y) +# define AV_GCC_VERSION_AT_LEAST(x,y) (__GNUC__ > (x) || __GNUC__ == (x) && __GNUC_MINOR__ >= (y)) +# define AV_GCC_VERSION_AT_MOST(x,y) (__GNUC__ < (x) || __GNUC__ == (x) && __GNUC_MINOR__ <= (y)) #else # define AV_GCC_VERSION_AT_LEAST(x,y) 0 +# define AV_GCC_VERSION_AT_MOST(x,y) 0 #endif +#ifndef av_always_inline #if AV_GCC_VERSION_AT_LEAST(3,1) # define av_always_inline __attribute__((always_inline)) inline #elif defined(_MSC_VER) @@ -39,6 +42,21 @@ #else # define av_always_inline inline #endif +#endif + +#ifndef av_extern_inline +#if defined(__ICL) && __ICL >= 1210 || defined(__GNUC_STDC_INLINE__) +# define av_extern_inline extern inline +#else +# define av_extern_inline inline +#endif +#endif + +#if AV_GCC_VERSION_AT_LEAST(3,4) +# define av_warn_unused_result __attribute__((warn_unused_result)) +#else +# define av_warn_unused_result +#endif #if AV_GCC_VERSION_AT_LEAST(3,1) # define av_noinline __attribute__((noinline)) @@ -80,6 +98,30 @@ # define attribute_deprecated #endif +/** + * Disable warnings about deprecated features + * This is useful for sections of code kept for backward compatibility and + * scheduled for removal. + */ +#ifndef AV_NOWARN_DEPRECATED +#if AV_GCC_VERSION_AT_LEAST(4,6) +# define AV_NOWARN_DEPRECATED(code) \ + _Pragma("GCC diagnostic push") \ + _Pragma("GCC diagnostic ignored \"-Wdeprecated-declarations\"") \ + code \ + _Pragma("GCC diagnostic pop") +#elif defined(_MSC_VER) +# define AV_NOWARN_DEPRECATED(code) \ + __pragma(warning(push)) \ + __pragma(warning(disable : 4996)) \ + code; \ + __pragma(warning(pop)) +#else +# define AV_NOWARN_DEPRECATED(code) code +#endif +#endif + + #if defined(__GNUC__) # define av_unused __attribute__((unused)) #else @@ -103,7 +145,7 @@ # define av_alias #endif -#if defined(__GNUC__) && !defined(__ICC) +#if defined(__GNUC__) && !defined(__INTEL_COMPILER) && !defined(__clang__) # define av_uninit(x) x=x #else # define av_uninit(x) x diff --git a/libavutil/audio_fifo.c b/libavutil/audio_fifo.c index b562537cf6..e4d38e0524 100644 --- a/libavutil/audio_fifo.c +++ b/libavutil/audio_fifo.c @@ -2,20 +2,20 @@ * Audio FIFO * Copyright (c) 2012 Justin Ruggles <justin.ruggles@gmail.com> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -51,7 +51,7 @@ void av_audio_fifo_free(AVAudioFifo *af) if (af->buf[i]) av_fifo_free(af->buf[i]); } - av_free(af->buf); + av_freep(&af->buf); } av_free(af); } @@ -76,7 +76,7 @@ AVAudioFifo *av_audio_fifo_alloc(enum AVSampleFormat sample_fmt, int channels, af->sample_size = buf_size / nb_samples; af->nb_buffers = av_sample_fmt_is_planar(sample_fmt) ? channels : 1; - af->buf = av_mallocz(af->nb_buffers * sizeof(*af->buf)); + af->buf = av_mallocz_array(af->nb_buffers, sizeof(*af->buf)); if (!af->buf) goto error; @@ -136,6 +136,49 @@ int av_audio_fifo_write(AVAudioFifo *af, void **data, int nb_samples) return nb_samples; } +int av_audio_fifo_peek(AVAudioFifo *af, void **data, int nb_samples) +{ + int i, ret, size; + + if (nb_samples < 0) + return AVERROR(EINVAL); + nb_samples = FFMIN(nb_samples, af->nb_samples); + if (!nb_samples) + return 0; + + size = nb_samples * af->sample_size; + for (i = 0; i < af->nb_buffers; i++) { + if ((ret = av_fifo_generic_peek(af->buf[i], data[i], size, NULL)) < 0) + return AVERROR_BUG; + } + + return nb_samples; +} + +int av_audio_fifo_peek_at(AVAudioFifo *af, void **data, int nb_samples, int offset) +{ + int i, ret, size; + + if (offset < 0 || offset >= af->nb_samples) + return AVERROR(EINVAL); + if (nb_samples < 0) + return AVERROR(EINVAL); + nb_samples = FFMIN(nb_samples, af->nb_samples); + if (!nb_samples) + return 0; + if (offset > af->nb_samples - nb_samples) + return AVERROR(EINVAL); + + offset *= af->sample_size; + size = nb_samples * af->sample_size; + for (i = 0; i < af->nb_buffers; i++) { + if ((ret = av_fifo_generic_peek_at(af->buf[i], data[i], offset, size, NULL)) < 0) + return AVERROR_BUG; + } + + return nb_samples; +} + int av_audio_fifo_read(AVAudioFifo *af, void **data, int nb_samples) { int i, ret, size; diff --git a/libavutil/audio_fifo.h b/libavutil/audio_fifo.h index e73123d864..d8a9194a8d 100644 --- a/libavutil/audio_fifo.h +++ b/libavutil/audio_fifo.h @@ -2,20 +2,20 @@ * Audio FIFO * Copyright (c) 2012 Justin Ruggles <justin.ruggles@gmail.com> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -73,6 +73,7 @@ AVAudioFifo *av_audio_fifo_alloc(enum AVSampleFormat sample_fmt, int channels, * @param nb_samples new allocation size, in samples * @return 0 if OK, or negative AVERROR code on failure */ +av_warn_unused_result int av_audio_fifo_realloc(AVAudioFifo *af, int nb_samples); /** @@ -88,11 +89,45 @@ int av_audio_fifo_realloc(AVAudioFifo *af, int nb_samples); * @param data audio data plane pointers * @param nb_samples number of samples to write * @return number of samples actually written, or negative AVERROR - * code on failure. + * code on failure. If successful, the number of samples + * actually written will always be nb_samples. */ int av_audio_fifo_write(AVAudioFifo *af, void **data, int nb_samples); /** + * Peek data from an AVAudioFifo. + * + * @see enum AVSampleFormat + * The documentation for AVSampleFormat describes the data layout. + * + * @param af AVAudioFifo to read from + * @param data audio data plane pointers + * @param nb_samples number of samples to peek + * @return number of samples actually peek, or negative AVERROR code + * on failure. The number of samples actually peek will not + * be greater than nb_samples, and will only be less than + * nb_samples if av_audio_fifo_size is less than nb_samples. + */ +int av_audio_fifo_peek(AVAudioFifo *af, void **data, int nb_samples); + +/** + * Peek data from an AVAudioFifo. + * + * @see enum AVSampleFormat + * The documentation for AVSampleFormat describes the data layout. + * + * @param af AVAudioFifo to read from + * @param data audio data plane pointers + * @param nb_samples number of samples to peek + * @param offset offset from current read position + * @return number of samples actually peek, or negative AVERROR code + * on failure. The number of samples actually peek will not + * be greater than nb_samples, and will only be less than + * nb_samples if av_audio_fifo_size is less than nb_samples. + */ +int av_audio_fifo_peek_at(AVAudioFifo *af, void **data, int nb_samples, int offset); + +/** * Read data from an AVAudioFifo. * * @see enum AVSampleFormat @@ -102,7 +137,9 @@ int av_audio_fifo_write(AVAudioFifo *af, void **data, int nb_samples); * @param data audio data plane pointers * @param nb_samples number of samples to read * @return number of samples actually read, or negative AVERROR code - * on failure. + * on failure. The number of samples actually read will not + * be greater than nb_samples, and will only be less than + * nb_samples if av_audio_fifo_size is less than nb_samples. */ int av_audio_fifo_read(AVAudioFifo *af, void **data, int nb_samples); diff --git a/libavutil/avassert.h b/libavutil/avassert.h index b223d26e8d..41f5e0eea7 100644 --- a/libavutil/avassert.h +++ b/libavutil/avassert.h @@ -1,20 +1,20 @@ /* * copyright (c) 2010 Michael Niedermayer <michaelni@gmx.at> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -36,7 +36,7 @@ */ #define av_assert0(cond) do { \ if (!(cond)) { \ - av_log(NULL, AV_LOG_FATAL, "Assertion %s failed at %s:%d\n", \ + av_log(NULL, AV_LOG_PANIC, "Assertion %s failed at %s:%d\n", \ AV_STRINGIFY(cond), __FILE__, __LINE__); \ abort(); \ } \ diff --git a/libavutil/avr32/bswap.h b/libavutil/avr32/bswap.h index 857f024054..e79d53f369 100644 --- a/libavutil/avr32/bswap.h +++ b/libavutil/avr32/bswap.h @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/avr32/intreadwrite.h b/libavutil/avr32/intreadwrite.h index e0049feb09..c6fd3aa470 100644 --- a/libavutil/avr32/intreadwrite.h +++ b/libavutil/avr32/intreadwrite.h @@ -1,20 +1,20 @@ /* * Copyright (c) 2009 Mans Rullgard <mans@mansr.com> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/avstring-test.c b/libavutil/avstring-test.c index 7732c22f0e..3e29661e96 100644 --- a/libavutil/avstring-test.c +++ b/libavutil/avstring-test.c @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -25,7 +25,8 @@ int main(void) { int i; - const char *strings[] = { + char *fullpath; + static const char * const strings[] = { "''", "", ":", @@ -65,5 +66,18 @@ int main(void) av_free(q); } + printf("Testing av_append_path_component()\n"); + #define TEST_APPEND_PATH_COMPONENT(path, component, expected) \ + fullpath = av_append_path_component((path), (component)); \ + printf("%s = %s\n", fullpath ? fullpath : "(null)", expected); \ + av_free(fullpath); + TEST_APPEND_PATH_COMPONENT(NULL, NULL, "(null)") + TEST_APPEND_PATH_COMPONENT("path", NULL, "path"); + TEST_APPEND_PATH_COMPONENT(NULL, "comp", "comp"); + TEST_APPEND_PATH_COMPONENT("path", "comp", "path/comp"); + TEST_APPEND_PATH_COMPONENT("path/", "comp", "path/comp"); + TEST_APPEND_PATH_COMPONENT("path", "/comp", "path/comp"); + TEST_APPEND_PATH_COMPONENT("path/", "/comp", "path/comp"); + TEST_APPEND_PATH_COMPONENT("path/path2/", "/comp/comp2", "path/path2/comp/comp2"); return 0; } diff --git a/libavutil/avstring.c b/libavutil/avstring.c index bc7d0bcb98..1787a1ef54 100644 --- a/libavutil/avstring.c +++ b/libavutil/avstring.c @@ -2,20 +2,20 @@ * Copyright (c) 2000, 2001, 2002 Fabrice Bellard * Copyright (c) 2007 Mans Rullgard * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -27,7 +27,9 @@ #include "config.h" #include "common.h" #include "mem.h" +#include "avassert.h" #include "avstring.h" +#include "bprint.h" int av_strstart(const char *str, const char *pfx, const char **ptr) { @@ -54,11 +56,11 @@ int av_stristart(const char *str, const char *pfx, const char **ptr) char *av_stristr(const char *s1, const char *s2) { if (!*s2) - return s1; + return (char*)(intptr_t)s1; do if (av_stristart(s1, s2, NULL)) - return s1; + return (char*)(intptr_t)s1; while (*s1++); return NULL; @@ -68,11 +70,11 @@ char *av_strnstr(const char *haystack, const char *needle, size_t hay_length) { size_t needle_len = strlen(needle); if (!needle_len) - return haystack; + return (char*)haystack; while (hay_length >= needle_len) { hay_length--; if (!memcmp(haystack, needle, needle_len)) - return haystack; + return (char*)haystack; haystack++; } return NULL; @@ -98,7 +100,7 @@ size_t av_strlcat(char *dst, const char *src, size_t size) size_t av_strlcatf(char *dst, size_t size, const char *fmt, ...) { - int len = strlen(dst); + size_t len = strlen(dst); va_list vl; va_start(vl, fmt); @@ -108,6 +110,32 @@ size_t av_strlcatf(char *dst, size_t size, const char *fmt, ...) return len; } +char *av_asprintf(const char *fmt, ...) +{ + char *p = NULL; + va_list va; + int len; + + va_start(va, fmt); + len = vsnprintf(NULL, 0, fmt, va); + va_end(va); + if (len < 0) + goto end; + + p = av_malloc(len + 1); + if (!p) + goto end; + + va_start(va, fmt); + len = vsnprintf(p, len + 1, fmt, va); + va_end(va); + if (len < 0) + av_freep(&p); + +end: + return p; +} + char *av_d2str(double d) { char *str = av_malloc(16); @@ -116,7 +144,7 @@ char *av_d2str(double d) return str; } -#define WHITESPACES " \n\t" +#define WHITESPACES " \n\t\r" char *av_get_token(const char **buf, const char *term) { @@ -153,6 +181,35 @@ char *av_get_token(const char **buf, const char *term) return ret; } +char *av_strtok(char *s, const char *delim, char **saveptr) +{ + char *tok; + + if (!s && !(s = *saveptr)) + return NULL; + + /* skip leading delimiters */ + s += strspn(s, delim); + + /* s now points to the first non delimiter char, or to the end of the string */ + if (!*s) { + *saveptr = NULL; + return NULL; + } + tok = s++; + + /* skip non delimiters */ + s += strcspn(s, delim); + if (*s) { + *s = 0; + *saveptr = s+1; + } else { + *saveptr = NULL; + } + + return tok; +} + int av_strcasecmp(const char *a, const char *b) { uint8_t c1, c2; @@ -212,6 +269,54 @@ const char *av_dirname(char *path) return path; } +char *av_append_path_component(const char *path, const char *component) +{ + size_t p_len, c_len; + char *fullpath; + + if (!path) + return av_strdup(component); + if (!component) + return av_strdup(path); + + p_len = strlen(path); + c_len = strlen(component); + if (p_len > SIZE_MAX - c_len || p_len + c_len > SIZE_MAX - 2) + return NULL; + fullpath = av_malloc(p_len + c_len + 2); + if (fullpath) { + if (p_len) { + av_strlcpy(fullpath, path, p_len + 1); + if (c_len) { + if (fullpath[p_len - 1] != '/' && component[0] != '/') + fullpath[p_len++] = '/'; + else if (fullpath[p_len - 1] == '/' && component[0] == '/') + p_len--; + } + } + av_strlcpy(&fullpath[p_len], component, c_len + 1); + fullpath[p_len + c_len] = 0; + } + return fullpath; +} + +int av_escape(char **dst, const char *src, const char *special_chars, + enum AVEscapeMode mode, int flags) +{ + AVBPrint dstbuf; + + av_bprint_init(&dstbuf, 1, AV_BPRINT_SIZE_UNLIMITED); + av_bprint_escape(&dstbuf, src, special_chars, mode, flags); + + if (!av_bprint_is_complete(&dstbuf)) { + av_bprint_finalize(&dstbuf, NULL); + return AVERROR(ENOMEM); + } else { + av_bprint_finalize(&dstbuf, dst); + return dstbuf.len; + } +} + int av_match_name(const char *name, const char *names) { const char *p; @@ -221,11 +326,112 @@ int av_match_name(const char *name, const char *names) return 0; namelen = strlen(name); - while ((p = strchr(names, ','))) { + while (*names) { + int negate = '-' == *names; + p = strchr(names, ','); + if (!p) + p = names + strlen(names); + names += negate; len = FFMAX(p - names, namelen); - if (!av_strncasecmp(name, names, len)) - return 1; - names = p + 1; + if (!av_strncasecmp(name, names, len) || !strncmp("ALL", names, FFMAX(3, p - names))) + return !negate; + names = p + (*p == ','); } - return !av_strcasecmp(name, names); + return 0; +} + +int av_utf8_decode(int32_t *codep, const uint8_t **bufp, const uint8_t *buf_end, + unsigned int flags) +{ + const uint8_t *p = *bufp; + uint32_t top; + uint64_t code; + int ret = 0, tail_len; + uint32_t overlong_encoding_mins[6] = { + 0x00000000, 0x00000080, 0x00000800, 0x00010000, 0x00200000, 0x04000000, + }; + + if (p >= buf_end) + return 0; + + code = *p++; + + /* first sequence byte starts with 10, or is 1111-1110 or 1111-1111, + which is not admitted */ + if ((code & 0xc0) == 0x80 || code >= 0xFE) { + ret = AVERROR(EILSEQ); + goto end; + } + top = (code & 128) >> 1; + + tail_len = 0; + while (code & top) { + int tmp; + tail_len++; + if (p >= buf_end) { + (*bufp) ++; + return AVERROR(EILSEQ); /* incomplete sequence */ + } + + /* we assume the byte to be in the form 10xx-xxxx */ + tmp = *p++ - 128; /* strip leading 1 */ + if (tmp>>6) { + (*bufp) ++; + return AVERROR(EILSEQ); + } + code = (code<<6) + tmp; + top <<= 5; + } + code &= (top << 1) - 1; + + /* check for overlong encodings */ + av_assert0(tail_len <= 5); + if (code < overlong_encoding_mins[tail_len]) { + ret = AVERROR(EILSEQ); + goto end; + } + + if (code >= 1U<<31) { + ret = AVERROR(EILSEQ); /* out-of-range value */ + goto end; + } + + *codep = code; + + if (code > 0x10FFFF && + !(flags & AV_UTF8_FLAG_ACCEPT_INVALID_BIG_CODES)) + ret = AVERROR(EILSEQ); + if (code < 0x20 && code != 0x9 && code != 0xA && code != 0xD && + flags & AV_UTF8_FLAG_EXCLUDE_XML_INVALID_CONTROL_CODES) + ret = AVERROR(EILSEQ); + if (code >= 0xD800 && code <= 0xDFFF && + !(flags & AV_UTF8_FLAG_ACCEPT_SURROGATES)) + ret = AVERROR(EILSEQ); + if ((code == 0xFFFE || code == 0xFFFF) && + !(flags & AV_UTF8_FLAG_ACCEPT_NON_CHARACTERS)) + ret = AVERROR(EILSEQ); + +end: + *bufp = p; + return ret; +} + +int av_match_list(const char *name, const char *list, char separator) +{ + const char *p, *q; + + for (p = name; p && *p; ) { + for (q = list; q && *q; ) { + int k; + for (k = 0; p[k] == q[k] || (p[k]*q[k] == 0 && p[k]+q[k] == separator); k++) + if (k && (!p[k] || p[k] == separator)) + return 1; + q = strchr(q, separator); + q += !!q; + } + p = strchr(p, separator); + p += !!p; + } + + return 0; } diff --git a/libavutil/avstring.h b/libavutil/avstring.h index 780f109164..dd2876990f 100644 --- a/libavutil/avstring.h +++ b/libavutil/avstring.h @@ -1,20 +1,20 @@ /* * Copyright (c) 2007 Mans Rullgard * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -22,6 +22,7 @@ #define AVUTIL_AVSTRING_H #include <stddef.h> +#include <stdint.h> #include "attributes.h" /** @@ -131,7 +132,31 @@ size_t av_strlcat(char *dst, const char *src, size_t size); size_t av_strlcatf(char *dst, size_t size, const char *fmt, ...) av_printf_format(3, 4); /** - * Convert a number to a av_malloced string. + * Get the count of continuous non zero chars starting from the beginning. + * + * @param len maximum number of characters to check in the string, that + * is the maximum value which is returned by the function + */ +static inline size_t av_strnlen(const char *s, size_t len) +{ + size_t i; + for (i = 0; i < len && s[i]; i++) + ; + return i; +} + +/** + * Print arguments following specified format into a large enough auto + * allocated buffer. It is similar to GNU asprintf(). + * @param fmt printf-compatible format string, specifying how the + * following parameters are used. + * @return the allocated string + * @note You have to free the string yourself with av_free(). + */ +char *av_asprintf(const char *fmt, ...) av_printf_format(1, 2); + +/** + * Convert a number to an av_malloced string. */ char *av_d2str(double d); @@ -152,6 +177,30 @@ char *av_d2str(double d); char *av_get_token(const char **buf, const char *term); /** + * Split the string into several tokens which can be accessed by + * successive calls to av_strtok(). + * + * A token is defined as a sequence of characters not belonging to the + * set specified in delim. + * + * On the first call to av_strtok(), s should point to the string to + * parse, and the value of saveptr is ignored. In subsequent calls, s + * should be NULL, and saveptr should be unchanged since the previous + * call. + * + * This function is similar to strtok_r() defined in POSIX.1. + * + * @param s the string to parse, may be NULL + * @param delim 0-terminated list of token delimiters, must be non-NULL + * @param saveptr user-provided pointer which points to stored + * information necessary for av_strtok() to continue scanning the same + * string. saveptr is updated to point to the next character after the + * first delimiter found, or to NULL if the string was terminated + * @return the found token, or NULL when no token is found + */ +char *av_strtok(char *s, const char *delim, char **saveptr); + +/** * Locale-independent conversion of ASCII isdigit. */ static inline av_const int av_isdigit(int c) @@ -205,7 +254,7 @@ static inline av_const int av_isxdigit(int c) return av_isdigit(c) || (c >= 'a' && c <= 'f'); } -/* +/** * Locale-independent case-insensitive compare. * @note This means only ASCII-range characters are case-insensitive */ @@ -233,9 +282,13 @@ const char *av_basename(const char *path); */ const char *av_dirname(char *path); - /** * Match instances of a name in a comma-separated list of names. + * List entries are checked from the start to the end of the names list, + * the first match ends further processing. If an entry prefixed with '-' + * matches, then 0 is returned. The "ALL" list entry is considered to + * match all names. + * * @param name Name to look for. * @param names List of names. * @return 1 on match, 0 otherwise. @@ -243,6 +296,106 @@ const char *av_dirname(char *path); int av_match_name(const char *name, const char *names); /** + * Append path component to the existing path. + * Path separator '/' is placed between when needed. + * Resulting string have to be freed with av_free(). + * @param path base path + * @param component component to be appended + * @return new path or NULL on error. + */ +char *av_append_path_component(const char *path, const char *component); + +enum AVEscapeMode { + AV_ESCAPE_MODE_AUTO, ///< Use auto-selected escaping mode. + AV_ESCAPE_MODE_BACKSLASH, ///< Use backslash escaping. + AV_ESCAPE_MODE_QUOTE, ///< Use single-quote escaping. +}; + +/** + * Consider spaces special and escape them even in the middle of the + * string. + * + * This is equivalent to adding the whitespace characters to the special + * characters lists, except it is guaranteed to use the exact same list + * of whitespace characters as the rest of libavutil. + */ +#define AV_ESCAPE_FLAG_WHITESPACE (1 << 0) + +/** + * Escape only specified special characters. + * Without this flag, escape also any characters that may be considered + * special by av_get_token(), such as the single quote. + */ +#define AV_ESCAPE_FLAG_STRICT (1 << 1) + +/** + * Escape string in src, and put the escaped string in an allocated + * string in *dst, which must be freed with av_free(). + * + * @param dst pointer where an allocated string is put + * @param src string to escape, must be non-NULL + * @param special_chars string containing the special characters which + * need to be escaped, can be NULL + * @param mode escape mode to employ, see AV_ESCAPE_MODE_* macros. + * Any unknown value for mode will be considered equivalent to + * AV_ESCAPE_MODE_BACKSLASH, but this behaviour can change without + * notice. + * @param flags flags which control how to escape, see AV_ESCAPE_FLAG_ macros + * @return the length of the allocated string, or a negative error code in case of error + * @see av_bprint_escape() + */ +av_warn_unused_result +int av_escape(char **dst, const char *src, const char *special_chars, + enum AVEscapeMode mode, int flags); + +#define AV_UTF8_FLAG_ACCEPT_INVALID_BIG_CODES 1 ///< accept codepoints over 0x10FFFF +#define AV_UTF8_FLAG_ACCEPT_NON_CHARACTERS 2 ///< accept non-characters - 0xFFFE and 0xFFFF +#define AV_UTF8_FLAG_ACCEPT_SURROGATES 4 ///< accept UTF-16 surrogates codes +#define AV_UTF8_FLAG_EXCLUDE_XML_INVALID_CONTROL_CODES 8 ///< exclude control codes not accepted by XML + +#define AV_UTF8_FLAG_ACCEPT_ALL \ + AV_UTF8_FLAG_ACCEPT_INVALID_BIG_CODES|AV_UTF8_FLAG_ACCEPT_NON_CHARACTERS|AV_UTF8_FLAG_ACCEPT_SURROGATES + +/** + * Read and decode a single UTF-8 code point (character) from the + * buffer in *buf, and update *buf to point to the next byte to + * decode. + * + * In case of an invalid byte sequence, the pointer will be updated to + * the next byte after the invalid sequence and the function will + * return an error code. + * + * Depending on the specified flags, the function will also fail in + * case the decoded code point does not belong to a valid range. + * + * @note For speed-relevant code a carefully implemented use of + * GET_UTF8() may be preferred. + * + * @param codep pointer used to return the parsed code in case of success. + * The value in *codep is set even in case the range check fails. + * @param bufp pointer to the address the first byte of the sequence + * to decode, updated by the function to point to the + * byte next after the decoded sequence + * @param buf_end pointer to the end of the buffer, points to the next + * byte past the last in the buffer. This is used to + * avoid buffer overreads (in case of an unfinished + * UTF-8 sequence towards the end of the buffer). + * @param flags a collection of AV_UTF8_FLAG_* flags + * @return >= 0 in case a sequence was successfully read, a negative + * value in case of invalid sequence + */ +av_warn_unused_result +int av_utf8_decode(int32_t *codep, const uint8_t **bufp, const uint8_t *buf_end, + unsigned int flags); + +/** + * Check if a name is in a list. + * @returns 0 if not found, or the 1 based index where it has been found in the + * list. + */ +int av_match_list(const char *name, const char *list, char separator); + +/** * @} */ diff --git a/libavutil/avutil.h b/libavutil/avutil.h index 688068eff8..9bcf674126 100644 --- a/libavutil/avutil.h +++ b/libavutil/avutil.h @@ -1,20 +1,20 @@ /* * copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -29,22 +29,23 @@ /** * @mainpage * - * @section libav_intro Introduction + * @section ffmpeg_intro Introduction * * This document describes the usage of the different libraries - * provided by Libav. + * provided by FFmpeg. * * @li @ref libavc "libavcodec" encoding/decoding library * @li @ref lavfi "libavfilter" graph-based frame editing library * @li @ref libavf "libavformat" I/O and muxing/demuxing library * @li @ref lavd "libavdevice" special devices muxing/demuxing library * @li @ref lavu "libavutil" common utility library - * @li @ref lavr "libavresample" audio resampling, format conversion and mixing - * @li @ref libsws "libswscale" color conversion and scaling library + * @li @ref lswr "libswresample" audio resampling, format conversion and mixing + * @li @ref lpp "libpostproc" post processing library + * @li @ref libsws "libswscale" color conversion and scaling library * - * @section libav_versioning Versioning and compatibility + * @section ffmpeg_versioning Versioning and compatibility * - * Each of the Libav libraries contains a version.h header, which defines a + * Each of the FFmpeg libraries contains a version.h header, which defines a * major, minor and micro version number with the * <em>LIBRARYNAME_VERSION_{MAJOR,MINOR,MICRO}</em> macros. The major version * number is incremented with backward incompatible changes - e.g. removing @@ -55,22 +56,22 @@ * might still want to check for - e.g. changing behavior in a previously * unspecified situation. * - * Libav guarantees backward API and ABI compatibility for each library as long + * FFmpeg guarantees backward API and ABI compatibility for each library as long * as its major version number is unchanged. This means that no public symbols * will be removed or renamed. Types and names of the public struct members and * values of public macros and enums will remain the same (unless they were * explicitly declared as not part of the public API). Documented behavior will * not change. * - * In other words, any correct program that works with a given Libav snapshot + * In other words, any correct program that works with a given FFmpeg snapshot * should work just as well without any changes with any later snapshot with the * same major versions. This applies to both rebuilding the program against new - * Libav versions or to replacing the dynamic Libav libraries that a program + * FFmpeg versions or to replacing the dynamic FFmpeg libraries that a program * links against. * * However, new public symbols may be added and new members may be appended to * public structs whose size is not part of public ABI (most public structs in - * Libav). New macros and enum values may be added. Behavior in undocumented + * FFmpeg). New macros and enum values may be added. Behavior in undocumented * situations may change slightly (and be documented). All those are accompanied * by an entry in doc/APIchanges and incrementing either the minor or micro * version number. @@ -80,7 +81,7 @@ * @defgroup lavu Common utility functions * * @brief - * libavutil contains the code shared across all the other Libav + * libavutil contains the code shared across all the other FFmpeg * libraries * * @note In order to use the functions provided by avutil you must include @@ -137,15 +138,13 @@ * * @{ * - * @defgroup lavu_internal Internal - * - * Not exported functions, for internal usage only + * @defgroup preproc_misc Preprocessor String Macros * * @{ * * @} * - * @defgroup preproc_misc Preprocessor String Macros + * @defgroup version_utils Library Version Macros * * @{ * @@ -200,6 +199,12 @@ enum AVMediaType { }; /** + * Return a string describing the media_type enum, NULL if media_type + * is unknown. + */ +const char *av_get_media_type_string(enum AVMediaType media_type); + +/** * @defgroup lavu_const Constants * @{ * @@ -220,7 +225,7 @@ enum AVMediaType { * @} * @defgroup lavu_time Timestamp specific * - * Libav internal timebase and timestamp definitions + * FFmpeg internal timebase and timestamp definitions * * @{ */ @@ -232,7 +237,7 @@ enum AVMediaType { * either pts or dts. */ -#define AV_NOPTS_VALUE INT64_C(0x8000000000000000) +#define AV_NOPTS_VALUE ((int64_t)UINT64_C(0x8000000000000000)) /** * Internal time base represented as integer @@ -257,7 +262,8 @@ enum AVMediaType { */ enum AVPictureType { - AV_PICTURE_TYPE_I = 1, ///< Intra + AV_PICTURE_TYPE_NONE = 0, ///< Undefined + AV_PICTURE_TYPE_I, ///< Intra AV_PICTURE_TYPE_P, ///< Predicted AV_PICTURE_TYPE_B, ///< Bi-dir predicted AV_PICTURE_TYPE_S, ///< S(GMC)-VOP MPEG4 @@ -279,10 +285,50 @@ char av_get_picture_type_char(enum AVPictureType pict_type); * @} */ +#include "common.h" #include "error.h" #include "rational.h" #include "version.h" #include "macros.h" +#include "mathematics.h" +#include "log.h" +#include "pixfmt.h" + +/** + * Return x default pointer in case p is NULL. + */ +static inline void *av_x_if_null(const void *p, const void *x) +{ + return (void *)(intptr_t)(p ? p : x); +} + +/** + * Compute the length of an integer list. + * + * @param elsize size in bytes of each list element (only 1, 2, 4 or 8) + * @param term list terminator (usually 0 or -1) + * @param list pointer to the list + * @return length of the list, in elements, not counting the terminator + */ +unsigned av_int_list_length_for_size(unsigned elsize, + const void *list, uint64_t term) av_pure; + +/** + * Compute the length of an integer list. + * + * @param term list terminator (usually 0 or -1) + * @param list pointer to the list + * @return length of the list, in elements, not counting the terminator + */ +#define av_int_list_length(list, term) \ + av_int_list_length_for_size(sizeof(*(list)), list, term) + +/** + * Open a file using a UTF-8 filename. + * The API of this function matches POSIX fopen(), errors are returned through + * errno. + */ +FILE *av_fopen_utf8(const char *path, const char *mode); /** * Return the fractional representation of the internal time base. diff --git a/libavutil/avutilres.rc b/libavutil/avutilres.rc new file mode 100644 index 0000000000..40a75eb377 --- /dev/null +++ b/libavutil/avutilres.rc @@ -0,0 +1,55 @@ +/* + * Windows resource file for libavutil + * + * Copyright (C) 2012 James Almer + * Copyright (C) 2013 Tiancheng "Timothy" Gu + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include <windows.h> +#include "libavutil/version.h" +#include "libavutil/ffversion.h" +#include "config.h" + +1 VERSIONINFO +FILEVERSION LIBAVUTIL_VERSION_MAJOR, LIBAVUTIL_VERSION_MINOR, LIBAVUTIL_VERSION_MICRO, 0 +PRODUCTVERSION LIBAVUTIL_VERSION_MAJOR, LIBAVUTIL_VERSION_MINOR, LIBAVUTIL_VERSION_MICRO, 0 +FILEFLAGSMASK VS_FFI_FILEFLAGSMASK +FILEOS VOS_NT_WINDOWS32 +FILETYPE VFT_DLL +{ + BLOCK "StringFileInfo" + { + BLOCK "040904B0" + { + VALUE "CompanyName", "FFmpeg Project" + VALUE "FileDescription", "FFmpeg utility library" + VALUE "FileVersion", AV_STRINGIFY(LIBAVUTIL_VERSION) + VALUE "InternalName", "libavutil" + VALUE "LegalCopyright", "Copyright (C) 2000-" AV_STRINGIFY(CONFIG_THIS_YEAR) " FFmpeg Project" + VALUE "OriginalFilename", "avutil" BUILDSUF "-" AV_STRINGIFY(LIBAVUTIL_VERSION_MAJOR) SLIBSUF + VALUE "ProductName", "FFmpeg" + VALUE "ProductVersion", FFMPEG_VERSION + } + } + + BLOCK "VarFileInfo" + { + VALUE "Translation", 0x0409, 0x04B0 + } +} diff --git a/libavutil/base64-test.c b/libavutil/base64-test.c index 3ce0518a84..19ecd2bee1 100644 --- a/libavutil/base64-test.c +++ b/libavutil/base64-test.c @@ -1,26 +1,28 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ +// LCOV_EXCL_START #include <stdint.h> #include <stdio.h> #include "common.h" #include "base64.h" +#include "timer.h" #define MAX_DATA_SIZE 1024 #define MAX_ENCODED_SIZE 2048 @@ -42,21 +44,40 @@ static int test_encode_decode(const uint8_t *data, unsigned int data_size, return 1; } - if ((data2_size = av_base64_decode(data2, encoded, max_data2_size)) < 0) { + if ((data2_size = av_base64_decode(data2, encoded, max_data2_size)) != data_size) { printf("Failed: cannot decode the encoded string\n" "Encoded:\n%s\n", encoded); return 1; } + if ((data2_size = av_base64_decode(data2, encoded, data_size)) != data_size) { + printf("Failed: cannot decode with minimal buffer\n" + "Encoded:\n%s\n", encoded); + return 1; + } if (memcmp(data2, data, data_size)) { printf("Failed: encoded/decoded data differs from original data\n"); return 1; } + if (av_base64_decode(NULL, encoded, 0) != 0) { + printf("Failed: decode to NULL buffer\n"); + return 1; + } + if (strlen(encoded)) { + char *end = strchr(encoded, '='); + if (!end) + end = encoded + strlen(encoded) - 1; + *end = '%'; + if (av_base64_decode(NULL, encoded, 0) >= 0) { + printf("Failed: error detection\n"); + return 1; + } + } printf("Passed!\n"); return 0; } -int main(void) +int main(int argc, char ** argv) { int i, error_count = 0; struct test { @@ -72,13 +93,36 @@ int main(void) { "666666", "NjY2NjY2"}, { "abc:def", "YWJjOmRlZg=="}, }; + char in[1024], out[2048]; printf("Encoding/decoding tests\n"); for (i = 0; i < FF_ARRAY_ELEMS(tests); i++) error_count += test_encode_decode(tests[i].data, strlen(tests[i].data), tests[i].encoded_ref); + if (argc>1 && !strcmp(argv[1], "-t")) { + memset(in, 123, sizeof(in)); + for(i=0; i<10000; i++){ + START_TIMER + av_base64_encode(out, sizeof(out), in, sizeof(in)); + STOP_TIMER("encode") + } + for(i=0; i<10000; i++){ + START_TIMER + av_base64_decode(in, out, sizeof(in)); + STOP_TIMER("decode") + } + + for(i=0; i<10000; i++){ + START_TIMER + av_base64_decode(NULL, out, 0); + STOP_TIMER("syntax check") + } + } + if (error_count) printf("Error Count: %d.\n", error_count); return !!error_count; } + +// LCOV_EXCL_STOP diff --git a/libavutil/base64.c b/libavutil/base64.c index 29677a646b..25ae8c411c 100644 --- a/libavutil/base64.c +++ b/libavutil/base64.c @@ -1,20 +1,20 @@ /* * Copyright (c) 2006 Ryan Martell. (rdm4@martellventures.com) * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -26,41 +26,107 @@ #include "common.h" #include "base64.h" +#include "intreadwrite.h" +#include "timer.h" /* ---------------- private code */ -static const uint8_t map2[] = +static const uint8_t map2[256] = { + 0xfe, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, + 0x3e, 0xff, 0xff, 0xff, 0x3f, 0x34, 0x35, 0x36, 0x37, 0x38, 0x39, 0x3a, 0x3b, 0x3c, 0x3d, 0xff, - 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x00, 0x01, + 0xff, 0xff, 0xfe, 0xff, 0xff, 0xff, 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f, 0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17, 0x18, 0x19, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f, 0x20, 0x21, 0x22, 0x23, 0x24, 0x25, 0x26, 0x27, 0x28, 0x29, 0x2a, 0x2b, - 0x2c, 0x2d, 0x2e, 0x2f, 0x30, 0x31, 0x32, 0x33 + 0x2c, 0x2d, 0x2e, 0x2f, 0x30, 0x31, 0x32, 0x33, + + 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, }; -int av_base64_decode(uint8_t *out, const char *in, int out_size) +#define BASE64_DEC_STEP(i) do { \ + bits = map2[in[i]]; \ + if (bits & 0x80) \ + goto out ## i; \ + v = i ? (v << 6) + bits : bits; \ +} while(0) + +int av_base64_decode(uint8_t *out, const char *in_str, int out_size) { - int i; - unsigned v = 0; uint8_t *dst = out; + uint8_t *end = out + out_size; + // no sign extension + const uint8_t *in = in_str; + unsigned bits = 0xff; + unsigned v; - for (i = 0; in[i] && in[i] != '='; i++) { - unsigned int index= in[i]-43; - if (index>=FF_ARRAY_ELEMS(map2) || map2[index] == 0xff) - return -1; - v = (v << 6) + map2[index]; - if (i & 3) { - if (dst - out < out_size) { - *dst++ = v >> (6 - 2 * (i & 3)); - } - } + while (end - dst > 3) { + BASE64_DEC_STEP(0); + BASE64_DEC_STEP(1); + BASE64_DEC_STEP(2); + BASE64_DEC_STEP(3); + // Using AV_WB32 directly confuses compiler + v = av_be2ne32(v << 8); + AV_WN32(dst, v); + dst += 3; + in += 4; + } + if (end - dst) { + BASE64_DEC_STEP(0); + BASE64_DEC_STEP(1); + BASE64_DEC_STEP(2); + BASE64_DEC_STEP(3); + *dst++ = v >> 16; + if (end - dst) + *dst++ = v >> 8; + if (end - dst) + *dst++ = v; + in += 4; + } + while (1) { + BASE64_DEC_STEP(0); + in++; + BASE64_DEC_STEP(0); + in++; + BASE64_DEC_STEP(0); + in++; + BASE64_DEC_STEP(0); + in++; } - return dst - out; +out3: + *dst++ = v >> 10; + v <<= 2; +out2: + *dst++ = v >> 4; +out1: +out0: + return bits & 1 ? AVERROR_INVALIDDATA : dst - out; } /***************************************************************************** @@ -82,15 +148,23 @@ char *av_base64_encode(char *out, int out_size, const uint8_t *in, int in_size) out_size < AV_BASE64_SIZE(in_size)) return NULL; ret = dst = out; + while (bytes_remaining > 3) { + i_bits = AV_RB32(in); + in += 3; bytes_remaining -= 3; + *dst++ = b64[ i_bits>>26 ]; + *dst++ = b64[(i_bits>>20) & 0x3F]; + *dst++ = b64[(i_bits>>14) & 0x3F]; + *dst++ = b64[(i_bits>>8 ) & 0x3F]; + } + i_bits = 0; while (bytes_remaining) { i_bits = (i_bits << 8) + *in++; bytes_remaining--; i_shift += 8; - - do { - *dst++ = b64[(i_bits << 6 >> i_shift) & 0x3f]; - i_shift -= 6; - } while (i_shift > 6 || (bytes_remaining == 0 && i_shift > 0)); + } + while (i_shift > 0) { + *dst++ = b64[(i_bits << 6 >> i_shift) & 0x3f]; + i_shift -= 6; } while ((dst - ret) & 3) *dst++ = '='; diff --git a/libavutil/base64.h b/libavutil/base64.h index 4750cf5c72..2954c12d42 100644 --- a/libavutil/base64.h +++ b/libavutil/base64.h @@ -1,20 +1,20 @@ /* * Copyright (c) 2006 Ryan Martell. (rdm4@martellventures.com) * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -29,32 +29,39 @@ * @{ */ - /** * Decode a base64-encoded string. * * @param out buffer for decoded data * @param in null-terminated input string * @param out_size size in bytes of the out buffer, must be at - * least 3/4 of the length of in + * least 3/4 of the length of in, that is AV_BASE64_DECODE_SIZE(strlen(in)) * @return number of bytes written, or a negative value in case of * invalid input */ int av_base64_decode(uint8_t *out, const char *in, int out_size); /** + * Calculate the output size in bytes needed to decode a base64 string + * with length x to a data buffer. + */ +#define AV_BASE64_DECODE_SIZE(x) ((x) * 3LL / 4) + +/** * Encode data to base64 and null-terminate. * * @param out buffer for encoded data - * @param out_size size in bytes of the output buffer, must be at - * least AV_BASE64_SIZE(in_size) - * @param in_size size in bytes of the 'in' buffer - * @return 'out' or NULL in case of error + * @param out_size size in bytes of the out buffer (including the + * null terminator), must be at least AV_BASE64_SIZE(in_size) + * @param in input buffer containing the data to encode + * @param in_size size in bytes of the in buffer + * @return out or NULL in case of error */ char *av_base64_encode(char *out, int out_size, const uint8_t *in, int in_size); /** - * Calculate the output size needed to base64-encode x bytes. + * Calculate the output size needed to base64-encode x bytes to a + * null-terminated string. */ #define AV_BASE64_SIZE(x) (((x)+2) / 3 * 4 + 1) diff --git a/libavutil/bfin/bswap.h b/libavutil/bfin/bswap.h index 2837a2f7e1..363ed40bc5 100644 --- a/libavutil/bfin/bswap.h +++ b/libavutil/bfin/bswap.h @@ -1,20 +1,20 @@ /* * Copyright (C) 2007 Marc Hoffman * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/bfin/timer.h b/libavutil/bfin/timer.h index 49d32a8f6b..644573daec 100644 --- a/libavutil/bfin/timer.h +++ b/libavutil/bfin/timer.h @@ -1,20 +1,20 @@ /* * Copyright (C) 2007 Marc Hoffman * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/blowfish-test.c b/libavutil/blowfish-test.c index 4281ca29bd..3ad7cd7665 100644 --- a/libavutil/blowfish-test.c +++ b/libavutil/blowfish-test.c @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/blowfish.c b/libavutil/blowfish.c index 87315766d7..abc0e03d4d 100644 --- a/libavutil/blowfish.c +++ b/libavutil/blowfish.c @@ -4,20 +4,20 @@ * * loosely based on Paul Kocher's implementation * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -27,15 +27,6 @@ #include "mem.h" #include "blowfish.h" -#if !FF_API_CRYPTO_CONTEXT -#define AV_BF_ROUNDS 16 - -struct AVBlowfish { - uint32_t p[AV_BF_ROUNDS + 2]; - uint32_t s[4][256]; -}; -#endif - static const uint32_t orig_p[AV_BF_ROUNDS + 2] = { 0x243F6A88, 0x85A308D3, 0x13198A2E, 0x03707344, 0xA4093822, 0x299F31D0, 0x082EFA98, 0xEC4E6C89, @@ -303,24 +294,12 @@ static const uint32_t orig_s[4][256] = { 0xB74E6132, 0xCE77E25B, 0x578FDFE3, 0x3AC372E6 } }; -static void F(AVBlowfish *ctx, uint32_t *xl, uint32_t *xr, int i) -{ - uint32_t Xl, Xr; - uint32_t y; - - Xl = *xl; - Xr = *xr; - - Xl ^= ctx->p[i]; - y = ctx->s[0][(Xl >> 24) & 0xFF]; - y += ctx->s[1][(Xl >> 16) & 0xFF]; - y ^= ctx->s[2][(Xl >> 8) & 0xFF]; - y += ctx->s[3][ Xl & 0xFF]; - Xr ^= y; - - *xl = Xr; - *xr = Xl; -} +#define F(Xl, Xr, P) \ + Xr ^=((( ctx->s[0][ Xl >> 24 ] \ + + ctx->s[1][(Xl >> 16) & 0xFF])\ + ^ ctx->s[2][(Xl >> 8) & 0xFF])\ + + ctx->s[3][ Xl & 0xFF])\ + ^ P; AVBlowfish *av_blowfish_alloc(void) { @@ -372,17 +351,21 @@ void av_blowfish_crypt_ecb(AVBlowfish *ctx, uint32_t *xl, uint32_t *xr, Xr = *xr; if (decrypt) { - for (i = AV_BF_ROUNDS + 1; i > 1; --i) - F(ctx, &Xl, &Xr, i); + Xl ^= ctx->p[AV_BF_ROUNDS + 1]; + for (i = AV_BF_ROUNDS; i > 0; i-=2) { + F(Xl, Xr, ctx->p[i ]); + F(Xr, Xl, ctx->p[i-1]); + } - Xl = Xl ^ ctx->p[1]; - Xr = Xr ^ ctx->p[0]; + Xr ^= ctx->p[0]; } else { - for (i = 0; i < AV_BF_ROUNDS; ++i) - F(ctx, &Xl, &Xr, i); + Xl ^= ctx->p[0]; + for (i = 1; i < AV_BF_ROUNDS+1; i+=2){ + F(Xl, Xr, ctx->p[i ]); + F(Xr, Xl, ctx->p[i+1]); + } - Xl = Xl ^ ctx->p[AV_BF_ROUNDS]; - Xr = Xr ^ ctx->p[AV_BF_ROUNDS + 1]; + Xr ^= ctx->p[AV_BF_ROUNDS + 1]; } *xl = Xr; diff --git a/libavutil/blowfish.h b/libavutil/blowfish.h index 4f86bf7cfb..9e289a40da 100644 --- a/libavutil/blowfish.h +++ b/libavutil/blowfish.h @@ -1,20 +1,21 @@ /* * Blowfish algorithm + * Copyright (c) 2012 Samuel Pitoiset * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -22,7 +23,6 @@ #define AVUTIL_BLOWFISH_H #include <stdint.h> -#include "version.h" /** * @defgroup lavu_blowfish Blowfish @@ -30,16 +30,12 @@ * @{ */ -#if FF_API_CRYPTO_CONTEXT #define AV_BF_ROUNDS 16 typedef struct AVBlowfish { uint32_t p[AV_BF_ROUNDS + 2]; uint32_t s[4][256]; } AVBlowfish; -#else -typedef struct AVBlowfish AVBlowfish; -#endif /** * Allocate an AVBlowfish context. diff --git a/libavutil/bprint-test.c b/libavutil/bprint-test.c new file mode 100644 index 0000000000..f081c4485e --- /dev/null +++ b/libavutil/bprint-test.c @@ -0,0 +1,93 @@ +/* + * Copyright (c) 2012 Nicolas George + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "bprint.c" + +#undef printf + +static void bprint_pascal(AVBPrint *b, unsigned size) +{ + unsigned i, j; + unsigned p[42]; + + av_assert0(size < FF_ARRAY_ELEMS(p)); + + p[0] = 1; + av_bprintf(b, "%8d\n", 1); + for (i = 1; i <= size; i++) { + p[i] = 1; + for (j = i - 1; j > 0; j--) + p[j] = p[j] + p[j - 1]; + for (j = 0; j <= i; j++) + av_bprintf(b, "%8d", p[j]); + av_bprintf(b, "\n"); + } +} + +int main(void) +{ + AVBPrint b; + char buf[256]; + struct tm testtime = { .tm_year = 100, .tm_mon = 11, .tm_mday = 20 }; + + av_bprint_init(&b, 0, -1); + bprint_pascal(&b, 5); + printf("Short text in unlimited buffer: %u/%u\n", (unsigned)strlen(b.str), b.len); + printf("%s\n", b.str); + av_bprint_finalize(&b, NULL); + + av_bprint_init(&b, 0, -1); + bprint_pascal(&b, 25); + printf("Long text in unlimited buffer: %u/%u\n", (unsigned)strlen(b.str), b.len); + av_bprint_finalize(&b, NULL); + + av_bprint_init(&b, 0, 2048); + bprint_pascal(&b, 25); + printf("Long text in limited buffer: %u/%u\n", (unsigned)strlen(b.str), b.len); + av_bprint_finalize(&b, NULL); + + av_bprint_init(&b, 0, 1); + bprint_pascal(&b, 5); + printf("Short text in automatic buffer: %u/%u\n", (unsigned)strlen(b.str), b.len); + + av_bprint_init(&b, 0, 1); + bprint_pascal(&b, 25); + printf("Long text in automatic buffer: %u/%u\n", (unsigned)strlen(b.str)/8*8, b.len); + /* Note that the size of the automatic buffer is arch-dependent. */ + + av_bprint_init(&b, 0, 0); + bprint_pascal(&b, 25); + printf("Long text count only buffer: %u/%u\n", (unsigned)strlen(b.str), b.len); + + av_bprint_init_for_buffer(&b, buf, sizeof(buf)); + bprint_pascal(&b, 25); + printf("Long text count only buffer: %u/%u\n", (unsigned)strlen(buf), b.len); + + av_bprint_init(&b, 0, -1); + av_bprint_strftime(&b, "%Y-%m-%d", &testtime); + printf("strftime full: %u/%u \"%s\"\n", (unsigned)strlen(buf), b.len, b.str); + av_bprint_finalize(&b, NULL); + + av_bprint_init(&b, 0, 8); + av_bprint_strftime(&b, "%Y-%m-%d", &testtime); + printf("strftime truncated: %u/%u \"%s\"\n", (unsigned)strlen(buf), b.len, b.str); + + return 0; +} diff --git a/libavutil/bprint.c b/libavutil/bprint.c new file mode 100644 index 0000000000..2f059c5ba6 --- /dev/null +++ b/libavutil/bprint.c @@ -0,0 +1,305 @@ +/* + * Copyright (c) 2012 Nicolas George + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include <stdarg.h> +#include <stdio.h> +#include <string.h> +#include <time.h> +#include "avassert.h" +#include "avstring.h" +#include "bprint.h" +#include "common.h" +#include "compat/va_copy.h" +#include "error.h" +#include "mem.h" + +#define av_bprint_room(buf) ((buf)->size - FFMIN((buf)->len, (buf)->size)) +#define av_bprint_is_allocated(buf) ((buf)->str != (buf)->reserved_internal_buffer) + +static int av_bprint_alloc(AVBPrint *buf, unsigned room) +{ + char *old_str, *new_str; + unsigned min_size, new_size; + + if (buf->size == buf->size_max) + return AVERROR(EIO); + if (!av_bprint_is_complete(buf)) + return AVERROR_INVALIDDATA; /* it is already truncated anyway */ + min_size = buf->len + 1 + FFMIN(UINT_MAX - buf->len - 1, room); + new_size = buf->size > buf->size_max / 2 ? buf->size_max : buf->size * 2; + if (new_size < min_size) + new_size = FFMIN(buf->size_max, min_size); + old_str = av_bprint_is_allocated(buf) ? buf->str : NULL; + new_str = av_realloc(old_str, new_size); + if (!new_str) + return AVERROR(ENOMEM); + if (!old_str) + memcpy(new_str, buf->str, buf->len + 1); + buf->str = new_str; + buf->size = new_size; + return 0; +} + +static void av_bprint_grow(AVBPrint *buf, unsigned extra_len) +{ + /* arbitrary margin to avoid small overflows */ + extra_len = FFMIN(extra_len, UINT_MAX - 5 - buf->len); + buf->len += extra_len; + if (buf->size) + buf->str[FFMIN(buf->len, buf->size - 1)] = 0; +} + +void av_bprint_init(AVBPrint *buf, unsigned size_init, unsigned size_max) +{ + unsigned size_auto = (char *)buf + sizeof(*buf) - + buf->reserved_internal_buffer; + + if (size_max == 1) + size_max = size_auto; + buf->str = buf->reserved_internal_buffer; + buf->len = 0; + buf->size = FFMIN(size_auto, size_max); + buf->size_max = size_max; + *buf->str = 0; + if (size_init > buf->size) + av_bprint_alloc(buf, size_init - 1); +} + +void av_bprint_init_for_buffer(AVBPrint *buf, char *buffer, unsigned size) +{ + buf->str = buffer; + buf->len = 0; + buf->size = size; + buf->size_max = size; + *buf->str = 0; +} + +void av_bprintf(AVBPrint *buf, const char *fmt, ...) +{ + unsigned room; + char *dst; + va_list vl; + int extra_len; + + while (1) { + room = av_bprint_room(buf); + dst = room ? buf->str + buf->len : NULL; + va_start(vl, fmt); + extra_len = vsnprintf(dst, room, fmt, vl); + va_end(vl); + if (extra_len <= 0) + return; + if (extra_len < room) + break; + if (av_bprint_alloc(buf, extra_len)) + break; + } + av_bprint_grow(buf, extra_len); +} + +void av_vbprintf(AVBPrint *buf, const char *fmt, va_list vl_arg) +{ + unsigned room; + char *dst; + int extra_len; + va_list vl; + + while (1) { + room = av_bprint_room(buf); + dst = room ? buf->str + buf->len : NULL; + va_copy(vl, vl_arg); + extra_len = vsnprintf(dst, room, fmt, vl); + va_end(vl); + if (extra_len <= 0) + return; + if (extra_len < room) + break; + if (av_bprint_alloc(buf, extra_len)) + break; + } + av_bprint_grow(buf, extra_len); +} + +void av_bprint_chars(AVBPrint *buf, char c, unsigned n) +{ + unsigned room, real_n; + + while (1) { + room = av_bprint_room(buf); + if (n < room) + break; + if (av_bprint_alloc(buf, n)) + break; + } + if (room) { + real_n = FFMIN(n, room - 1); + memset(buf->str + buf->len, c, real_n); + } + av_bprint_grow(buf, n); +} + +void av_bprint_append_data(AVBPrint *buf, const char *data, unsigned size) +{ + unsigned room, real_n; + + while (1) { + room = av_bprint_room(buf); + if (size < room) + break; + if (av_bprint_alloc(buf, size)) + break; + } + if (room) { + real_n = FFMIN(size, room - 1); + memcpy(buf->str + buf->len, data, real_n); + } + av_bprint_grow(buf, size); +} + +void av_bprint_strftime(AVBPrint *buf, const char *fmt, const struct tm *tm) +{ + unsigned room; + size_t l; + + if (!*fmt) + return; + while (1) { + room = av_bprint_room(buf); + if (room && (l = strftime(buf->str + buf->len, room, fmt, tm))) + break; + /* strftime does not tell us how much room it would need: let us + retry with twice as much until the buffer is large enough */ + room = !room ? strlen(fmt) + 1 : + room <= INT_MAX / 2 ? room * 2 : INT_MAX; + if (av_bprint_alloc(buf, room)) { + /* impossible to grow, try to manage something useful anyway */ + room = av_bprint_room(buf); + if (room < 1024) { + /* if strftime fails because the buffer has (almost) reached + its maximum size, let us try in a local buffer; 1k should + be enough to format any real date+time string */ + char buf2[1024]; + if ((l = strftime(buf2, sizeof(buf2), fmt, tm))) { + av_bprintf(buf, "%s", buf2); + return; + } + } + if (room) { + /* if anything else failed and the buffer is not already + truncated, let us add a stock string and force truncation */ + static const char txt[] = "[truncated strftime output]"; + memset(buf->str + buf->len, '!', room); + memcpy(buf->str + buf->len, txt, FFMIN(sizeof(txt) - 1, room)); + av_bprint_grow(buf, room); /* force truncation */ + } + return; + } + } + av_bprint_grow(buf, l); +} + +void av_bprint_get_buffer(AVBPrint *buf, unsigned size, + unsigned char **mem, unsigned *actual_size) +{ + if (size > av_bprint_room(buf)) + av_bprint_alloc(buf, size); + *actual_size = av_bprint_room(buf); + *mem = *actual_size ? buf->str + buf->len : NULL; +} + +void av_bprint_clear(AVBPrint *buf) +{ + if (buf->len) { + *buf->str = 0; + buf->len = 0; + } +} + +int av_bprint_finalize(AVBPrint *buf, char **ret_str) +{ + unsigned real_size = FFMIN(buf->len + 1, buf->size); + char *str; + int ret = 0; + + if (ret_str) { + if (av_bprint_is_allocated(buf)) { + str = av_realloc(buf->str, real_size); + if (!str) + str = buf->str; + buf->str = NULL; + } else { + str = av_malloc(real_size); + if (str) + memcpy(str, buf->str, real_size); + else + ret = AVERROR(ENOMEM); + } + *ret_str = str; + } else { + if (av_bprint_is_allocated(buf)) + av_freep(&buf->str); + } + buf->size = real_size; + return ret; +} + +#define WHITESPACES " \n\t\r" + +void av_bprint_escape(AVBPrint *dstbuf, const char *src, const char *special_chars, + enum AVEscapeMode mode, int flags) +{ + const char *src0 = src; + + if (mode == AV_ESCAPE_MODE_AUTO) + mode = AV_ESCAPE_MODE_BACKSLASH; /* TODO: implement a heuristic */ + + switch (mode) { + case AV_ESCAPE_MODE_QUOTE: + /* enclose the string between '' */ + av_bprint_chars(dstbuf, '\'', 1); + for (; *src; src++) { + if (*src == '\'') + av_bprintf(dstbuf, "'\\''"); + else + av_bprint_chars(dstbuf, *src, 1); + } + av_bprint_chars(dstbuf, '\'', 1); + break; + + /* case AV_ESCAPE_MODE_BACKSLASH or unknown mode */ + default: + /* \-escape characters */ + for (; *src; src++) { + int is_first_last = src == src0 || !*(src+1); + int is_ws = !!strchr(WHITESPACES, *src); + int is_strictly_special = special_chars && strchr(special_chars, *src); + int is_special = + is_strictly_special || strchr("'\\", *src) || + (is_ws && (flags & AV_ESCAPE_FLAG_WHITESPACE)); + + if (is_strictly_special || + (!(flags & AV_ESCAPE_FLAG_STRICT) && + (is_special || (is_ws && is_first_last)))) + av_bprint_chars(dstbuf, '\\', 1); + av_bprint_chars(dstbuf, *src, 1); + } + break; + } +} diff --git a/libavutil/bprint.h b/libavutil/bprint.h new file mode 100644 index 0000000000..c09b1ac1e1 --- /dev/null +++ b/libavutil/bprint.h @@ -0,0 +1,219 @@ +/* + * Copyright (c) 2012 Nicolas George + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#ifndef AVUTIL_BPRINT_H +#define AVUTIL_BPRINT_H + +#include <stdarg.h> + +#include "attributes.h" +#include "avstring.h" + +/** + * Define a structure with extra padding to a fixed size + * This helps ensuring binary compatibility with future versions. + */ + +#define FF_PAD_STRUCTURE(name, size, ...) \ +struct ff_pad_helper_##name { __VA_ARGS__ }; \ +typedef struct name { \ + __VA_ARGS__ \ + char reserved_padding[size - sizeof(struct ff_pad_helper_##name)]; \ +} name; + +/** + * Buffer to print data progressively + * + * The string buffer grows as necessary and is always 0-terminated. + * The content of the string is never accessed, and thus is + * encoding-agnostic and can even hold binary data. + * + * Small buffers are kept in the structure itself, and thus require no + * memory allocation at all (unless the contents of the buffer is needed + * after the structure goes out of scope). This is almost as lightweight as + * declaring a local "char buf[512]". + * + * The length of the string can go beyond the allocated size: the buffer is + * then truncated, but the functions still keep account of the actual total + * length. + * + * In other words, buf->len can be greater than buf->size and records the + * total length of what would have been to the buffer if there had been + * enough memory. + * + * Append operations do not need to be tested for failure: if a memory + * allocation fails, data stop being appended to the buffer, but the length + * is still updated. This situation can be tested with + * av_bprint_is_complete(). + * + * The size_max field determines several possible behaviours: + * + * size_max = -1 (= UINT_MAX) or any large value will let the buffer be + * reallocated as necessary, with an amortized linear cost. + * + * size_max = 0 prevents writing anything to the buffer: only the total + * length is computed. The write operations can then possibly be repeated in + * a buffer with exactly the necessary size + * (using size_init = size_max = len + 1). + * + * size_max = 1 is automatically replaced by the exact size available in the + * structure itself, thus ensuring no dynamic memory allocation. The + * internal buffer is large enough to hold a reasonable paragraph of text, + * such as the current paragraph. + */ + +FF_PAD_STRUCTURE(AVBPrint, 1024, + char *str; /**< string so far */ + unsigned len; /**< length so far */ + unsigned size; /**< allocated memory */ + unsigned size_max; /**< maximum allocated memory */ + char reserved_internal_buffer[1]; +) + +/** + * Convenience macros for special values for av_bprint_init() size_max + * parameter. + */ +#define AV_BPRINT_SIZE_UNLIMITED ((unsigned)-1) +#define AV_BPRINT_SIZE_AUTOMATIC 1 +#define AV_BPRINT_SIZE_COUNT_ONLY 0 + +/** + * Init a print buffer. + * + * @param buf buffer to init + * @param size_init initial size (including the final 0) + * @param size_max maximum size; + * 0 means do not write anything, just count the length; + * 1 is replaced by the maximum value for automatic storage; + * any large value means that the internal buffer will be + * reallocated as needed up to that limit; -1 is converted to + * UINT_MAX, the largest limit possible. + * Check also AV_BPRINT_SIZE_* macros. + */ +void av_bprint_init(AVBPrint *buf, unsigned size_init, unsigned size_max); + +/** + * Init a print buffer using a pre-existing buffer. + * + * The buffer will not be reallocated. + * + * @param buf buffer structure to init + * @param buffer byte buffer to use for the string data + * @param size size of buffer + */ +void av_bprint_init_for_buffer(AVBPrint *buf, char *buffer, unsigned size); + +/** + * Append a formatted string to a print buffer. + */ +void av_bprintf(AVBPrint *buf, const char *fmt, ...) av_printf_format(2, 3); + +/** + * Append a formatted string to a print buffer. + */ +void av_vbprintf(AVBPrint *buf, const char *fmt, va_list vl_arg); + +/** + * Append char c n times to a print buffer. + */ +void av_bprint_chars(AVBPrint *buf, char c, unsigned n); + +/** + * Append data to a print buffer. + * + * param buf bprint buffer to use + * param data pointer to data + * param size size of data + */ +void av_bprint_append_data(AVBPrint *buf, const char *data, unsigned size); + +struct tm; +/** + * Append a formatted date and time to a print buffer. + * + * param buf bprint buffer to use + * param fmt date and time format string, see strftime() + * param tm broken-down time structure to translate + * + * @note due to poor design of the standard strftime function, it may + * produce poor results if the format string expands to a very long text and + * the bprint buffer is near the limit stated by the size_max option. + */ +void av_bprint_strftime(AVBPrint *buf, const char *fmt, const struct tm *tm); + +/** + * Allocate bytes in the buffer for external use. + * + * @param[in] buf buffer structure + * @param[in] size required size + * @param[out] mem pointer to the memory area + * @param[out] actual_size size of the memory area after allocation; + * can be larger or smaller than size + */ +void av_bprint_get_buffer(AVBPrint *buf, unsigned size, + unsigned char **mem, unsigned *actual_size); + +/** + * Reset the string to "" but keep internal allocated data. + */ +void av_bprint_clear(AVBPrint *buf); + +/** + * Test if the print buffer is complete (not truncated). + * + * It may have been truncated due to a memory allocation failure + * or the size_max limit (compare size and size_max if necessary). + */ +static inline int av_bprint_is_complete(const AVBPrint *buf) +{ + return buf->len < buf->size; +} + +/** + * Finalize a print buffer. + * + * The print buffer can no longer be used afterwards, + * but the len and size fields are still valid. + * + * @arg[out] ret_str if not NULL, used to return a permanent copy of the + * buffer contents, or NULL if memory allocation fails; + * if NULL, the buffer is discarded and freed + * @return 0 for success or error code (probably AVERROR(ENOMEM)) + */ +int av_bprint_finalize(AVBPrint *buf, char **ret_str); + +/** + * Escape the content in src and append it to dstbuf. + * + * @param dstbuf already inited destination bprint buffer + * @param src string containing the text to escape + * @param special_chars string containing the special characters which + * need to be escaped, can be NULL + * @param mode escape mode to employ, see AV_ESCAPE_MODE_* macros. + * Any unknown value for mode will be considered equivalent to + * AV_ESCAPE_MODE_BACKSLASH, but this behaviour can change without + * notice. + * @param flags flags which control how to escape, see AV_ESCAPE_FLAG_* macros + */ +void av_bprint_escape(AVBPrint *dstbuf, const char *src, const char *special_chars, + enum AVEscapeMode mode, int flags); + +#endif /* AVUTIL_BPRINT_H */ diff --git a/libavutil/bswap.h b/libavutil/bswap.h index 93a6016b8c..91cb79538d 100644 --- a/libavutil/bswap.h +++ b/libavutil/bswap.h @@ -1,20 +1,20 @@ /* * copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -40,8 +40,6 @@ # include "arm/bswap.h" #elif ARCH_AVR32 # include "avr32/bswap.h" -#elif ARCH_BFIN -# include "bfin/bswap.h" #elif ARCH_SH4 # include "sh4/bswap.h" #elif ARCH_X86 diff --git a/libavutil/buffer.c b/libavutil/buffer.c index 6681002345..694e116a3c 100644 --- a/libavutil/buffer.c +++ b/libavutil/buffer.c @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -103,14 +103,17 @@ AVBufferRef *av_buffer_ref(AVBufferRef *buf) return ret; } -void av_buffer_unref(AVBufferRef **buf) +static void buffer_replace(AVBufferRef **dst, AVBufferRef **src) { AVBuffer *b; - if (!buf || !*buf) - return; - b = (*buf)->buffer; - av_freep(buf); + b = (*dst)->buffer; + + if (src) { + **dst = **src; + av_freep(src); + } else + av_freep(dst); if (!avpriv_atomic_int_add_and_fetch(&b->refcount, -1)) { b->free(b->opaque, b->data); @@ -118,12 +121,30 @@ void av_buffer_unref(AVBufferRef **buf) } } +void av_buffer_unref(AVBufferRef **buf) +{ + if (!buf || !*buf) + return; + + buffer_replace(buf, NULL); +} + int av_buffer_is_writable(const AVBufferRef *buf) { if (buf->buffer->flags & AV_BUFFER_FLAG_READONLY) return 0; - return avpriv_atomic_int_add_and_fetch(&buf->buffer->refcount, 0) == 1; + return avpriv_atomic_int_get(&buf->buffer->refcount) == 1; +} + +void *av_buffer_get_opaque(const AVBufferRef *buf) +{ + return buf->buffer->opaque; +} + +int av_buffer_get_ref_count(const AVBufferRef *buf) +{ + return buf->buffer->refcount; } int av_buffer_make_writable(AVBufferRef **pbuf) @@ -138,8 +159,8 @@ int av_buffer_make_writable(AVBufferRef **pbuf) return AVERROR(ENOMEM); memcpy(newbuf->data, buf->data, buf->size); - av_buffer_unref(pbuf); - *pbuf = newbuf; + + buffer_replace(pbuf, &newbuf); return 0; } @@ -180,8 +201,7 @@ int av_buffer_realloc(AVBufferRef **pbuf, int size) memcpy(new->data, buf->data, FFMIN(size, buf->size)); - av_buffer_unref(pbuf); - *pbuf = new; + buffer_replace(pbuf, &new); return 0; } @@ -264,15 +284,60 @@ void av_buffer_pool_uninit(AVBufferPool **ppool) buffer_pool_free(pool); } +#if USE_ATOMICS +/* remove the whole buffer list from the pool and return it */ +static BufferPoolEntry *get_pool(AVBufferPool *pool) +{ + BufferPoolEntry *cur = *(void * volatile *)&pool->pool, *last = NULL; + + while (cur != last) { + last = cur; + cur = avpriv_atomic_ptr_cas((void * volatile *)&pool->pool, last, NULL); + if (!cur) + return NULL; + } + + return cur; +} + +static void add_to_pool(BufferPoolEntry *buf) +{ + AVBufferPool *pool; + BufferPoolEntry *cur, *end = buf; + + if (!buf) + return; + pool = buf->pool; + + while (end->next) + end = end->next; + + while (avpriv_atomic_ptr_cas((void * volatile *)&pool->pool, NULL, buf)) { + /* pool is not empty, retrieve it and append it to our list */ + cur = get_pool(pool); + end->next = cur; + while (end->next) + end = end->next; + } +} +#endif + static void pool_release_buffer(void *opaque, uint8_t *data) { BufferPoolEntry *buf = opaque; AVBufferPool *pool = buf->pool; + if(CONFIG_MEMORY_POISONING) + memset(buf->data, FF_MEMORY_POISON, pool->size); + +#if USE_ATOMICS + add_to_pool(buf); +#else ff_mutex_lock(&pool->mutex); buf->next = pool->pool; pool->pool = buf; ff_mutex_unlock(&pool->mutex); +#endif if (!avpriv_atomic_int_add_and_fetch(&pool->refcount, -1)) buffer_pool_free(pool); @@ -304,6 +369,11 @@ static AVBufferRef *pool_alloc_buffer(AVBufferPool *pool) ret->buffer->opaque = buf; ret->buffer->free = pool_release_buffer; +#if USE_ATOMICS + avpriv_atomic_int_add_and_fetch(&pool->refcount, 1); + avpriv_atomic_int_add_and_fetch(&pool->nb_allocated, 1); +#endif + return ret; } @@ -312,6 +382,29 @@ AVBufferRef *av_buffer_pool_get(AVBufferPool *pool) AVBufferRef *ret; BufferPoolEntry *buf; +#if USE_ATOMICS + /* check whether the pool is empty */ + buf = get_pool(pool); + if (!buf && pool->refcount <= pool->nb_allocated) { + av_log(NULL, AV_LOG_DEBUG, "Pool race dectected, spining to avoid overallocation and eventual OOM\n"); + while (!buf && avpriv_atomic_int_get(&pool->refcount) <= avpriv_atomic_int_get(&pool->nb_allocated)) + buf = get_pool(pool); + } + + if (!buf) + return pool_alloc_buffer(pool); + + /* keep the first entry, return the rest of the list to the pool */ + add_to_pool(buf->next); + buf->next = NULL; + + ret = av_buffer_create(buf->data, pool->size, pool_release_buffer, + buf, 0); + if (!ret) { + add_to_pool(buf); + return NULL; + } +#else ff_mutex_lock(&pool->mutex); buf = pool->pool; if (buf) { @@ -325,6 +418,7 @@ AVBufferRef *av_buffer_pool_get(AVBufferPool *pool) ret = pool_alloc_buffer(pool); } ff_mutex_unlock(&pool->mutex); +#endif if (ret) avpriv_atomic_int_add_and_fetch(&pool->refcount, 1); diff --git a/libavutil/buffer.h b/libavutil/buffer.h index 46779cae9b..0c0ce12cf2 100644 --- a/libavutil/buffer.h +++ b/libavutil/buffer.h @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -49,7 +49,7 @@ * Use av_buffer_unref() to free a reference (this will automatically free the * data once all the references are freed). * - * The convention throughout this API and the rest of Libav is such that the + * The convention throughout this API and the rest of FFmpeg is such that the * buffer is considered writable if there exists only one reference to it (and * it has not been marked as read-only). The av_buffer_is_writable() function is * provided to check whether this is true and av_buffer_make_writable() will @@ -122,7 +122,7 @@ AVBufferRef *av_buffer_allocz(int size); * @param data data array * @param size size of data in bytes * @param free a callback for freeing this buffer's data - * @param opaque parameter to be passed to free + * @param opaque parameter to be got for processing or passed to free * @param flags a combination of AV_BUFFER_FLAG_* * * @return an AVBufferRef referring to data on success, NULL on failure. @@ -163,6 +163,13 @@ void av_buffer_unref(AVBufferRef **buf); int av_buffer_is_writable(const AVBufferRef *buf); /** + * @return the opaque parameter set by av_buffer_create. + */ +void *av_buffer_get_opaque(const AVBufferRef *buf); + +int av_buffer_get_ref_count(const AVBufferRef *buf); + +/** * Create a writable reference from a given buffer reference, avoiding data copy * if possible. * diff --git a/libavutil/buffer_internal.h b/libavutil/buffer_internal.h index 64344d8262..29ce8a643c 100644 --- a/libavutil/buffer_internal.h +++ b/libavutil/buffer_internal.h @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -87,6 +87,8 @@ struct AVBufferPool { */ volatile int refcount; + volatile int nb_allocated; + int size; void *opaque; AVBufferRef* (*alloc)(int size); diff --git a/libavutil/camellia-test.c b/libavutil/camellia-test.c new file mode 100644 index 0000000000..ad90587345 --- /dev/null +++ b/libavutil/camellia-test.c @@ -0,0 +1,78 @@ +/* + * An implementation of the CAMELLIA algorithm as mentioned in RFC3713 + * Copyright (c) 2014 Supraja Meedinti + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "camellia.c" +#include "log.h" + +#include <stdio.h> +#include <stdlib.h> + +int main(int argc, char *argv[]) +{ + const uint8_t Key[3][32] = { + {0x01, 0x23, 0x45, 0x67, 0x89, 0xab, 0xcd, 0xef, 0xfe, 0xdc, 0xba, 0x98, 0x76, 0x54, 0x32, 0x10}, + {0x01, 0x23, 0x45, 0x67, 0x89, 0xab, 0xcd, 0xef, 0xfe, 0xdc, 0xba, 0x98, 0x76, 0x54, 0x32, 0x10, 0x00, 0x11, 0x22, 0x33, 0x44, 0x55, 0x66, 0x77}, + {0x01, 0x23, 0x45, 0x67, 0x89, 0xab, 0xcd, 0xef, 0xfe, 0xdc, 0xba, 0x98, 0x76, 0x54, 0x32, 0x10, 0x00, 0x11, 0x22, 0x33, 0x44, 0x55, 0x66, 0x77, 0x88, 0x99, 0xaa, 0xbb, 0xcc, 0xdd, 0xee, 0xff} + }; + const uint8_t rct[3][16] = { + {0x67, 0x67, 0x31, 0x38, 0x54, 0x96, 0x69, 0x73, 0x08, 0x57, 0x06, 0x56, 0x48, 0xea, 0xbe, 0x43}, + {0xb4, 0x99, 0x34, 0x01, 0xb3, 0xe9,0x96, 0xf8, 0x4e, 0xe5, 0xce, 0xe7, 0xd7, 0x9b, 0x09, 0xb9}, + {0x9a, 0xcc, 0x23, 0x7d, 0xff, 0x16, 0xd7, 0x6c, 0x20, 0xef, 0x7c, 0x91, 0x9e, 0x3a, 0x75, 0x09} + }; + const uint8_t rpt[32] = {0x01, 0x23, 0x45, 0x67, 0x89, 0xab, 0xcd, 0xef, 0xfe, 0xdc, 0xba, 0x98, 0x76, 0x54, 0x32, 0x10, 0x01, 0x23, 0x45, 0x67, 0x89, 0xab, 0xcd, 0xef, 0xfe, 0xdc, 0xba, 0x98, 0x76, 0x54, 0x32, 0x10}; + const int kbits[3] = {128, 192, 256}; + int i, j, err = 0; + uint8_t temp[32], iv[16]; + AVCAMELLIA *cs; + cs = av_camellia_alloc(); + if (!cs) + return 1; + for (j = 0; j < 3; j++) { + av_camellia_init(cs, Key[j], kbits[j]); + av_camellia_crypt(cs, temp, rpt, 1, NULL, 0); + for (i = 0; i < 16; i++) { + if (rct[j][i] != temp[i]) { + av_log(NULL, AV_LOG_ERROR, "%d %02x %02x\n", i, rct[j][i], temp[i]); + err = 1; + } + } + av_camellia_crypt(cs, temp, rct[j], 1, NULL, 1); + for (i = 0; i < 16; i++) { + if (rpt[i] != temp[i]) { + av_log(NULL, AV_LOG_ERROR, "%d %02x %02x\n", i, rpt[i], temp[i]); + err = 1; + } + } + } + av_camellia_init(cs, Key[0], 128); + memcpy(iv, "HALLO123HALLO123", 16); + av_camellia_crypt(cs, temp, rpt, 2, iv, 0); + memcpy(iv, "HALLO123HALLO123", 16); + av_camellia_crypt(cs, temp, temp, 2, iv, 1); + for (i = 0; i < 32; i++) { + if (rpt[i] != temp[i]) { + av_log(NULL, AV_LOG_ERROR, "%d %02x %02x\n", i, rpt[i], temp[i]); + err = 1; + } + } + av_free(cs); + return err; +} diff --git a/libavutil/camellia.c b/libavutil/camellia.c new file mode 100644 index 0000000000..f33ee9babc --- /dev/null +++ b/libavutil/camellia.c @@ -0,0 +1,412 @@ +/* + * An implementation of the CAMELLIA algorithm as mentioned in RFC3713 + * Copyright (c) 2014 Supraja Meedinti + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ +#include "camellia.h" +#include "common.h" +#include "intreadwrite.h" +#include "attributes.h" + +#define LR32(x,c) ((x) << (c) | (x) >> (32 - (c))) +#define RR32(x,c) ((x) >> (c) | (x) << (32 - (c))) + +#define MASK8 0xff +#define MASK32 0xffffffff +#define MASK64 0xffffffffffffffff + +#define Sigma1 0xA09E667F3BCC908B +#define Sigma2 0xB67AE8584CAA73B2 +#define Sigma3 0xC6EF372FE94F82BE +#define Sigma4 0x54FF53A5F1D36F1C +#define Sigma5 0x10E527FADE682D1D +#define Sigma6 0xB05688C2B3E6C1FD + +static uint64_t SP[8][256]; + +typedef struct AVCAMELLIA { + uint64_t Kw[4]; + uint64_t Ke[6]; + uint64_t K[24]; + int key_bits; +} AVCAMELLIA; + +static const uint8_t SBOX1[256] = { +112, 130, 44, 236, 179, 39, 192, 229, 228, 133, 87, 53, 234, 12, 174, 65, + 35, 239, 107, 147, 69, 25, 165, 33, 237, 14, 79, 78, 29, 101, 146, 189, +134, 184, 175, 143, 124, 235, 31, 206, 62, 48, 220, 95, 94, 197, 11, 26, +166, 225, 57, 202, 213, 71, 93, 61, 217, 1, 90, 214, 81, 86, 108, 77, +139, 13, 154, 102, 251, 204, 176, 45, 116, 18, 43, 32, 240, 177, 132, 153, +223, 76, 203, 194, 52, 126, 118, 5, 109, 183, 169, 49, 209, 23, 4, 215, + 20, 88, 58, 97, 222, 27, 17, 28, 50, 15, 156, 22, 83, 24, 242, 34, +254, 68, 207, 178, 195, 181, 122, 145, 36, 8, 232, 168, 96, 252, 105, 80, +170, 208, 160, 125, 161, 137, 98, 151, 84, 91, 30, 149, 224, 255, 100, 210, + 16, 196, 0, 72, 163, 247, 117, 219, 138, 3, 230, 218, 9, 63, 221, 148, +135, 92, 131, 2, 205, 74, 144, 51, 115, 103, 246, 243, 157, 127, 191, 226, + 82, 155, 216, 38, 200, 55, 198, 59, 129, 150, 111, 75, 19, 190, 99, 46, +233, 121, 167, 140, 159, 110, 188, 142, 41, 245, 249, 182, 47, 253, 180, 89, +120, 152, 6, 106, 231, 70, 113, 186, 212, 37, 171, 66, 136, 162, 141, 250, +114, 7, 185, 85, 248, 238, 172, 10, 54, 73, 42, 104, 60, 56, 241, 164, + 64, 40, 211, 123, 187, 201, 67, 193, 21, 227, 173, 244, 119, 199, 128, 158 +}; + +static const uint8_t SBOX2[256] = { +224, 5, 88, 217, 103, 78, 129, 203, 201, 11, 174, 106, 213, 24, 93, 130, + 70, 223, 214, 39, 138, 50, 75, 66, 219, 28, 158, 156, 58, 202, 37, 123, + 13, 113, 95, 31, 248, 215, 62, 157, 124, 96, 185, 190, 188, 139, 22, 52, + 77, 195, 114, 149, 171, 142, 186, 122, 179, 2, 180, 173, 162, 172, 216, 154, + 23, 26, 53, 204, 247, 153, 97, 90, 232, 36, 86, 64, 225, 99, 9, 51, +191, 152, 151, 133, 104, 252, 236, 10, 218, 111, 83, 98, 163, 46, 8, 175, + 40, 176, 116, 194, 189, 54, 34, 56, 100, 30, 57, 44, 166, 48, 229, 68, +253, 136, 159, 101, 135, 107, 244, 35, 72, 16, 209, 81, 192, 249, 210, 160, + 85, 161, 65, 250, 67, 19, 196, 47, 168, 182, 60, 43, 193, 255, 200, 165, + 32, 137, 0, 144, 71, 239, 234, 183, 21, 6, 205, 181, 18, 126, 187, 41, + 15, 184, 7, 4, 155, 148, 33, 102, 230, 206, 237, 231, 59, 254, 127, 197, +164, 55, 177, 76, 145, 110, 141, 118, 3, 45, 222, 150, 38, 125, 198, 92, +211, 242, 79, 25, 63, 220, 121, 29, 82, 235, 243, 109, 94, 251, 105, 178, +240, 49, 12, 212, 207, 140, 226, 117, 169, 74, 87, 132, 17, 69, 27, 245, +228, 14, 115, 170, 241, 221, 89, 20, 108, 146, 84, 208, 120, 112, 227, 73, +128, 80, 167, 246, 119, 147, 134, 131, 42, 199, 91, 233, 238, 143, 1, 61 +}; + +static const uint8_t SBOX3[256] = { + 56, 65, 22, 118, 217, 147, 96, 242, 114, 194, 171, 154, 117, 6, 87, 160, +145, 247, 181, 201, 162, 140, 210, 144, 246, 7, 167, 39, 142, 178, 73, 222, + 67, 92, 215, 199, 62, 245, 143, 103, 31, 24, 110, 175, 47, 226, 133, 13, + 83, 240, 156, 101, 234, 163, 174, 158, 236, 128, 45, 107, 168, 43, 54, 166, +197, 134, 77, 51, 253, 102, 88, 150, 58, 9, 149, 16, 120, 216, 66, 204, +239, 38, 229, 97, 26, 63, 59, 130, 182, 219, 212, 152, 232, 139, 2, 235, + 10, 44, 29, 176, 111, 141, 136, 14, 25, 135, 78, 11, 169, 12, 121, 17, +127, 34, 231, 89, 225, 218, 61, 200, 18, 4, 116, 84, 48, 126, 180, 40, + 85, 104, 80, 190, 208, 196, 49, 203, 42, 173, 15, 202, 112, 255, 50, 105, + 8, 98, 0, 36, 209, 251, 186, 237, 69, 129, 115, 109, 132, 159, 238, 74, +195, 46, 193, 1, 230, 37, 72, 153, 185, 179, 123, 249, 206, 191, 223, 113, + 41, 205, 108, 19, 100, 155, 99, 157, 192, 75, 183, 165, 137, 95, 177, 23, +244, 188, 211, 70, 207, 55, 94, 71, 148, 250, 252, 91, 151, 254, 90, 172, + 60, 76, 3, 53, 243, 35, 184, 93, 106, 146, 213, 33, 68, 81, 198, 125, + 57, 131, 220, 170, 124, 119, 86, 5, 27, 164, 21, 52, 30, 28, 248, 82, + 32, 20, 233, 189, 221, 228, 161, 224, 138, 241, 214, 122, 187, 227, 64, 79 +}; + +static const uint8_t SBOX4[256] = { +112, 44, 179, 192, 228, 87, 234, 174, 35, 107, 69, 165, 237, 79, 29, 146, +134, 175, 124, 31, 62, 220, 94, 11, 166, 57, 213, 93, 217, 90, 81, 108, +139, 154, 251, 176, 116, 43, 240, 132, 223, 203, 52, 118, 109, 169, 209, 4, + 20, 58, 222, 17, 50, 156, 83, 242, 254, 207, 195, 122, 36, 232, 96, 105, +170, 160, 161, 98, 84, 30, 224, 100, 16, 0, 163, 117, 138, 230, 9, 221, +135, 131, 205, 144, 115, 246, 157, 191, 82, 216, 200, 198, 129, 111, 19, 99, +233, 167, 159, 188, 41, 249, 47, 180, 120, 6, 231, 113, 212, 171, 136, 141, +114, 185, 248, 172, 54, 42, 60, 241, 64, 211, 187, 67, 21, 173, 119, 128, +130, 236, 39, 229, 133, 53, 12, 65, 239, 147, 25, 33, 14, 78, 101, 189, +184, 143, 235, 206, 48, 95, 197, 26, 225, 202, 71, 61, 1, 214, 86, 77, + 13, 102, 204, 45, 18, 32, 177, 153, 76, 194, 126, 5, 183, 49, 23, 215, + 88, 97, 27, 28, 15, 22, 24, 34, 68, 178, 181, 145, 8, 168, 252, 80, +208, 125, 137, 151, 91, 149, 255, 210, 196, 72, 247, 219, 3, 218, 63, 148, + 92, 2, 74, 51, 103, 243, 127, 226, 155, 38, 55, 59, 150, 75, 190, 46, +121, 140, 110, 142, 245, 182, 253, 89, 152, 106, 70, 186, 37, 66, 162, 250, + 7, 85, 238, 10, 73, 104, 56, 164, 40, 123, 201, 193, 227, 244, 199, 158 +}; + +const int av_camellia_size = sizeof(AVCAMELLIA); + +static void LR128(uint64_t d[2], const uint64_t K[2], int x) +{ + int i = 0; + if (64 <= x && x < 128) { + i = 1; + x -= 64; + } + if (x <= 0 || x >= 128) { + d[0] = K[i]; + d[1] = K[!i]; + return; + } + d[0] = (K[i] << x | K[!i] >> (64 - x)); + d[1] = (K[!i] << x | K[i] >> (64 - x)); +} + +static uint64_t F(uint64_t F_IN, uint64_t KE) +{ + KE ^= F_IN; + F_IN=SP[0][KE >> 56]^SP[1][(KE >> 48) & MASK8]^SP[2][(KE >> 40) & MASK8]^SP[3][(KE >> 32) & MASK8]^SP[4][(KE >> 24) & MASK8]^SP[5][(KE >> 16) & MASK8]^SP[6][(KE >> 8) & MASK8]^SP[7][KE & MASK8]; + return F_IN; +} + +static uint64_t FL(uint64_t FL_IN, uint64_t KE) +{ + uint32_t x1, x2, k1, k2; + x1 = FL_IN >> 32; + x2 = FL_IN & MASK32; + k1 = KE >> 32; + k2 = KE & MASK32; + x2 = x2 ^ LR32((x1 & k1), 1); + x1 = x1 ^ (x2 | k2); + return ((uint64_t)x1 << 32) | (uint64_t)x2; +} + +static uint64_t FLINV(uint64_t FLINV_IN, uint64_t KE) +{ + uint32_t x1, x2, k1, k2; + x1 = FLINV_IN >> 32; + x2 = FLINV_IN & MASK32; + k1 = KE >> 32; + k2 = KE & MASK32; + x1 = x1 ^ (x2 | k2); + x2 = x2 ^ LR32((x1 & k1), 1); + return ((uint64_t)x1 << 32) | (uint64_t)x2; +} + +static const uint8_t shifts[2][12] = { + {0, 15, 15, 45, 45, 60, 94, 94, 111}, + {0, 15, 15, 30, 45, 45, 60, 60, 77, 94, 94, 111} +}; + +static const uint8_t vars[2][12] = { + {2, 0, 2, 0, 2, 2, 0, 2, 0}, + {3, 1, 2, 3, 0, 2, 1, 3, 0, 1, 2, 0} +}; + +static void generate_round_keys(AVCAMELLIA *cs, uint64_t Kl[2], uint64_t Kr[2], uint64_t Ka[2], uint64_t Kb[2]) +{ + int i; + uint64_t *Kd[4], d[2]; + Kd[0] = Kl; + Kd[1] = Kr; + Kd[2] = Ka; + Kd[3] = Kb; + cs->Kw[0] = Kl[0]; + cs->Kw[1] = Kl[1]; + if (cs->key_bits == 128) { + for (i = 0; i < 9; i++) { + LR128(d, Kd[vars[0][i]], shifts[0][i]); + cs->K[2*i] = d[0]; + cs->K[2*i+1] = d[1]; + } + LR128(d, Kd[0], 60); + cs->K[9] = d[1]; + LR128(d, Kd[2], 30); + cs->Ke[0] = d[0]; + cs->Ke[1] = d[1]; + LR128(d, Kd[0], 77); + cs->Ke[2] = d[0]; + cs->Ke[3] = d[1]; + LR128(d, Kd[2], 111); + cs->Kw[2] = d[0]; + cs->Kw[3] = d[1]; + } else { + for (i = 0; i < 12; i++) { + LR128(d, Kd[vars[1][i]], shifts[1][i]); + cs->K[2*i] = d[0]; + cs->K[2*i+1] = d[1]; + } + LR128(d, Kd[1], 30); + cs->Ke[0] = d[0]; + cs->Ke[1] = d[1]; + LR128(d, Kd[0], 60); + cs->Ke[2] = d[0]; + cs->Ke[3] = d[1]; + LR128(d, Kd[2], 77); + cs->Ke[4] = d[0]; + cs->Ke[5] = d[1]; + LR128(d, Kd[3], 111); + cs->Kw[2] = d[0]; + cs->Kw[3] = d[1]; + } +} + +static void camellia_encrypt(AVCAMELLIA *cs, uint8_t *dst, const uint8_t *src) +{ + uint64_t D1, D2; + D1 = AV_RB64(src); + D2 = AV_RB64(src + 8); + D1 ^= cs->Kw[0]; + D2 ^= cs->Kw[1]; + D2 ^= F(D1, cs->K[0]); + D1 ^= F(D2, cs->K[1]); + D2 ^= F(D1, cs->K[2]); + D1 ^= F(D2, cs->K[3]); + D2 ^= F(D1, cs->K[4]); + D1 ^= F(D2, cs->K[5]); + D1 = FL(D1, cs->Ke[0]); + D2 = FLINV(D2, cs->Ke[1]); + D2 ^= F(D1, cs->K[6]); + D1 ^= F(D2, cs->K[7]); + D2 ^= F(D1, cs->K[8]); + D1 ^= F(D2, cs->K[9]); + D2 ^= F(D1, cs->K[10]); + D1 ^= F(D2, cs->K[11]); + D1 = FL(D1, cs->Ke[2]); + D2 = FLINV(D2, cs->Ke[3]); + D2 ^= F(D1, cs->K[12]); + D1 ^= F(D2, cs->K[13]); + D2 ^= F(D1, cs->K[14]); + D1 ^= F(D2, cs->K[15]); + D2 ^= F(D1, cs->K[16]); + D1 ^= F(D2, cs->K[17]); + if (cs->key_bits != 128) { + D1 = FL(D1, cs->Ke[4]); + D2 = FLINV(D2, cs->Ke[5]); + D2 ^= F(D1, cs->K[18]); + D1 ^= F(D2, cs->K[19]); + D2 ^= F(D1, cs->K[20]); + D1 ^= F(D2, cs->K[21]); + D2 ^= F(D1, cs->K[22]); + D1 ^= F(D2, cs->K[23]); + } + D2 ^= cs->Kw[2]; + D1 ^= cs->Kw[3]; + AV_WB64(dst, D2); + AV_WB64(dst + 8, D1); +} + +static void camellia_decrypt(AVCAMELLIA *cs, uint8_t *dst, const uint8_t *src, uint8_t *iv) +{ + uint64_t D1, D2; + D1 = AV_RB64(src); + D2 = AV_RB64(src + 8); + D1 ^= cs->Kw[2]; + D2 ^= cs->Kw[3]; + if (cs->key_bits != 128) { + D2 ^= F(D1, cs->K[23]); + D1 ^= F(D2, cs->K[22]); + D2 ^= F(D1, cs->K[21]); + D1 ^= F(D2, cs->K[20]); + D2 ^= F(D1, cs->K[19]); + D1 ^= F(D2, cs->K[18]); + D1 = FL(D1, cs->Ke[5]); + D2 = FLINV(D2, cs->Ke[4]); + } + D2 ^= F(D1, cs->K[17]); + D1 ^= F(D2, cs->K[16]); + D2 ^= F(D1, cs->K[15]); + D1 ^= F(D2, cs->K[14]); + D2 ^= F(D1, cs->K[13]); + D1 ^= F(D2, cs->K[12]); + D1 = FL(D1, cs->Ke[3]); + D2 = FLINV(D2, cs->Ke[2]); + D2 ^= F(D1, cs->K[11]); + D1 ^= F(D2, cs->K[10]); + D2 ^= F(D1, cs->K[9]); + D1 ^= F(D2, cs->K[8]); + D2 ^= F(D1, cs->K[7]); + D1 ^= F(D2, cs->K[6]); + D1 = FL(D1, cs->Ke[1]); + D2 = FLINV(D2, cs->Ke[0]); + D2 ^= F(D1, cs->K[5]); + D1 ^= F(D2, cs->K[4]); + D2 ^= F(D1, cs->K[3]); + D1 ^= F(D2, cs->K[2]); + D2 ^= F(D1, cs->K[1]); + D1 ^= F(D2, cs->K[0]); + D2 ^= cs->Kw[0]; + D1 ^= cs->Kw[1]; + if (iv) { + D2 ^= AV_RB64(iv); + D1 ^= AV_RB64(iv + 8); + memcpy(iv, src, 16); + } + AV_WB64(dst, D2); + AV_WB64(dst + 8, D1); +} + +static void computeSP(void) +{ + uint64_t z; + int i; + for (i = 0; i < 256; i++) { + z = SBOX1[i]; + SP[0][i] = (z << 56) ^ (z << 48) ^ (z << 40) ^ (z << 24) ^ z; + SP[7][i] = (z << 56) ^ (z << 48) ^ (z << 40) ^ (z << 24) ^ (z << 16) ^ (z << 8); + z = SBOX2[i]; + SP[1][i] = (z << 48) ^ (z << 40) ^ (z << 32) ^ (z << 24) ^ (z << 16); + SP[4][i] = (z << 48) ^ (z << 40) ^ (z << 32) ^ (z << 16) ^ (z << 8) ^ z; + z = SBOX3[i]; + SP[2][i] = (z << 56) ^ (z << 40) ^ (z << 32) ^ (z << 16) ^ (z << 8); + SP[5][i] = (z << 56) ^ (z << 40) ^ (z << 32) ^ (z << 24) ^ (z << 8) ^ z; + z = SBOX4[i]; + SP[3][i] = (z << 56) ^ (z << 48) ^ (z << 32) ^ (z << 8) ^ z; + SP[6][i] = (z << 56) ^ (z << 48) ^ (z << 32) ^ (z << 24) ^ (z << 16) ^ z; + } +} + +struct AVCAMELLIA *av_camellia_alloc(void) +{ + return av_mallocz(sizeof(struct AVCAMELLIA)); +} + +av_cold int av_camellia_init(AVCAMELLIA *cs, const uint8_t *key, int key_bits) +{ + uint64_t Kl[2], Kr[2], Ka[2], Kb[2]; + uint64_t D1, D2; + if (key_bits != 128 && key_bits != 192 && key_bits != 256) + return AVERROR(EINVAL); + memset(Kb, 0, sizeof(Kb)); + memset(Kr, 0, sizeof(Kr)); + cs->key_bits = key_bits; + Kl[0] = AV_RB64(key); + Kl[1] = AV_RB64(key + 8); + if (key_bits == 192) { + Kr[0] = AV_RB64(key + 16); + Kr[1] = ~Kr[0]; + } else if (key_bits == 256) { + Kr[0] = AV_RB64(key + 16); + Kr[1] = AV_RB64(key + 24); + } + computeSP(); + D1 = Kl[0] ^ Kr[0]; + D2 = Kl[1] ^ Kr[1]; + D2 ^= F(D1, Sigma1); + D1 ^= F(D2, Sigma2); + D1 ^= Kl[0]; + D2 ^= Kl[1]; + D2 ^= F(D1, Sigma3); + D1 ^= F(D2, Sigma4); + Ka[0] = D1; + Ka[1] = D2; + if (key_bits != 128) { + D1 = Ka[0] ^ Kr[0]; + D2 = Ka[1] ^ Kr[1]; + D2 ^= F(D1, Sigma5); + D1 ^= F(D2, Sigma6); + Kb[0] = D1; + Kb[1] = D2; + } + generate_round_keys(cs, Kl, Kr, Ka, Kb); + return 0; +} + +void av_camellia_crypt(AVCAMELLIA *cs, uint8_t *dst, const uint8_t *src, int count, uint8_t *iv, int decrypt) +{ + int i; + while (count--) { + if (decrypt) { + camellia_decrypt(cs, dst, src, iv); + } else { + if (iv) { + for (i = 0; i < 16; i++) + dst[i] = src[i] ^ iv[i]; + camellia_encrypt(cs, dst, dst); + memcpy(iv, dst, 16); + } else { + camellia_encrypt(cs, dst, src); + } + } + src = src + 16; + dst = dst + 16; + } +} diff --git a/libavutil/camellia.h b/libavutil/camellia.h new file mode 100644 index 0000000000..e674c9b9a4 --- /dev/null +++ b/libavutil/camellia.h @@ -0,0 +1,70 @@ +/* + * An implementation of the CAMELLIA algorithm as mentioned in RFC3713 + * Copyright (c) 2014 Supraja Meedinti + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#ifndef AVUTIL_CAMELLIA_H +#define AVUTIL_CAMELLIA_H + +#include <stdint.h> + + +/** + * @file + * @brief Public header for libavutil CAMELLIA algorithm + * @defgroup lavu_camellia CAMELLIA + * @ingroup lavu_crypto + * @{ + */ + +extern const int av_camellia_size; + +struct AVCAMELLIA; + +/** + * Allocate an AVCAMELLIA context + * To free the struct: av_free(ptr) + */ +struct AVCAMELLIA *av_camellia_alloc(void); + +/** + * Initialize an AVCAMELLIA context. + * + * @param ctx an AVCAMELLIA context + * @param key a key of 16, 24, 32 bytes used for encryption/decryption + * @param key_bits number of keybits: possible are 128, 192, 256 + */ +int av_camellia_init(struct AVCAMELLIA *ctx, const uint8_t *key, int key_bits); + +/** + * Encrypt or decrypt a buffer using a previously initialized context + * + * @param ctx an AVCAMELLIA context + * @param dst destination array, can be equal to src + * @param src source array, can be equal to dst + * @param count number of 16 byte blocks + * @paran iv initialization vector for CBC mode, NULL for ECB mode + * @param decrypt 0 for encryption, 1 for decryption + */ +void av_camellia_crypt(struct AVCAMELLIA *ctx, uint8_t *dst, const uint8_t *src, int count, uint8_t* iv, int decrypt); + +/** + * @} + */ +#endif /* AVUTIL_CAMELLIA_H */ diff --git a/libavutil/cast5-test.c b/libavutil/cast5-test.c new file mode 100644 index 0000000000..3776d11d67 --- /dev/null +++ b/libavutil/cast5-test.c @@ -0,0 +1,106 @@ +/* + * An implementation of the CAST128 algorithm as mentioned in RFC2144 + * Copyright (c) 2014 Supraja Meedinti + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "cast5.c" +#include "log.h" + +#include <stdio.h> +#include <stdlib.h> + +int main(int argc, char** argv) +{ + + static const uint8_t Key[3][16] = { + {0x01, 0x23, 0x45, 0x67, 0x12, 0x34, 0x56, 0x78, 0x23, 0x45, 0x67, 0x89, 0x34, 0x56, 0x78, 0x9a}, + {0x01, 0x23, 0x45, 0x67, 0x12, 0x34, 0x56, 0x78, 0x23, 0x45}, + {0x01, 0x23, 0x45, 0x67, 0x12} + }; + static const uint8_t rpt[8] = {0x01, 0x23, 0x45, 0x67, 0x89, 0xab, 0xcd, 0xef}; + static const uint8_t rct[3][8] = { + {0x23, 0x8b, 0x4f, 0xe5, 0x84, 0x7e, 0x44, 0xb2}, + {0xeb, 0x6a, 0x71, 0x1a, 0x2c, 0x02, 0x27, 0x1b}, + {0x7a, 0xc8, 0x16, 0xd1, 0x6e, 0x9b, 0x30, 0x2e} + }; + static const uint8_t rct2[2][16] = { + {0xee, 0xa9, 0xd0, 0xa2, 0x49, 0xfd, 0x3b, 0xa6, 0xb3, 0x43, 0x6f, 0xb8, 0x9d, 0x6d, 0xca, 0x92}, + {0xb2, 0xc9, 0x5e, 0xb0, 0x0c, 0x31, 0xad, 0x71, 0x80, 0xac, 0x05, 0xb8, 0xe8, 0x3d, 0x69, 0x6e} + }; + static const uint8_t iv[8] = {0xee, 0xa9, 0xd0, 0xa2, 0x49, 0xfd, 0x3b, 0xa6}; + static uint8_t rpt2[2][16]; + int i, j, err = 0; + static const int key_bits[3] = {128, 80, 40}; + uint8_t temp[8]; + AVCAST5 *cs; + cs = av_cast5_alloc(); + if (!cs) + return 1; + for (j = 0; j < 3; j++){ + + av_cast5_init(cs, Key[j], key_bits[j]); + av_cast5_crypt(cs, temp, rpt, 1, 0); + for (i = 0;i < 8; i++){ + if (rct[j][i] != temp[i]){ + av_log(NULL, AV_LOG_ERROR, "%d %02x %02x\n", i, rct[j][i], temp[i]); + err = 1; + } + } + + av_cast5_crypt(cs, temp, rct[j], 1, 1); + for (i =0; i < 8; i++) { + if (rpt[i] != temp[i]) { + av_log(NULL, AV_LOG_ERROR, "%d %02x %02x\n", i, rpt[i], temp[i]); + err = 1; + } + } + } + memcpy(rpt2[0], Key[0], 16); + memcpy(rpt2[1], Key[0], 16); + for (i = 0; i < 1000000; i++){ + av_cast5_init(cs, rpt2[1], 128); + av_cast5_crypt(cs, rpt2[0], rpt2[0], 2, 0); + av_cast5_init(cs, rpt2[0], 128); + av_cast5_crypt(cs, rpt2[1], rpt2[1], 2, 0); + } + for (j = 0; j < 2; j++) { + for (i = 0; i < 16; i++) { + if (rct2[j][i] != rpt2[j][i]) { + av_log(NULL, AV_LOG_ERROR, "%d %02x %02x\n", i, rct2[j][i], rpt2[j][i]); + err = 1; + } + } + } + for (j = 0; j < 3; j++) { + + av_cast5_init(cs, Key[j], key_bits[j]); + memcpy(temp, iv, 8); + av_cast5_crypt2(cs, rpt2[0], rct2[0], 2, temp, 0); + memcpy(temp, iv, 8); + av_cast5_crypt2(cs, rpt2[0], rpt2[0], 2, temp, 1); + for (i = 0; i < 16; i++) { + if (rct2[0][i] != rpt2[0][i]) { + av_log(NULL, AV_LOG_ERROR, "%d %02x %02x\n", i, rct2[0][i], rpt2[0][i]); + err = 1; + } + } + } + av_free(cs); + return err; +} diff --git a/libavutil/cast5.c b/libavutil/cast5.c new file mode 100644 index 0000000000..445eb55c12 --- /dev/null +++ b/libavutil/cast5.c @@ -0,0 +1,507 @@ +/* + * An implementation of the CAST128 algorithm as mentioned in RFC2144 + * Copyright (c) 2014 Supraja Meedinti + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ +#include "cast5.h" +#include "common.h" +#include "intreadwrite.h" +#include "attributes.h" + +#define IA(x) ((x) >> 24) +#define IB(x) (((x) >> 16) & 0xff) +#define IC(x) (((x) >> 8) & 0xff) +#define ID(x) ((x) & 0xff) + +#define LR(x, c) (((x) << (c)) | ((x) >> (32 - (c)))) + +#define F3(l, r, i) \ + do { \ + I = LR(cs->Km[i] - r, cs->Kr[i]); \ + f = ((S1[IA(I)] + S2[IB(I)]) ^ S3[IC(I)]) - S4[ID(I)]; \ + l = f ^ l; \ + } while (0) + +#define F2(l, r, i) \ + do { \ + I = LR(cs->Km[i] ^ r, cs->Kr[i]); \ + f = ((S1[IA(I)] - S2[IB(I)]) + S3[IC(I)]) ^ S4[ID(I)]; \ + l = f ^ l; \ + } while (0) + +#define F1(l, r, i) \ + do { \ + I = LR(cs->Km[i] + r, cs->Kr[i]); \ + f = ((S1[IA(I)] ^ S2[IB(I)]) - S3[IC(I)]) + S4[ID(I)]; \ + l = f ^ l; \ + } while (0) + +#define COMPUTE_Z \ + do { \ + z[0] = x[0] ^ S5[IB(x[3])] ^ S6[ID(x[3])] ^ S7[IA(x[3])] ^ S8[IC(x[3])] ^ S7[IA(x[2])]; \ + z[1] = x[2] ^ S5[IA(z[0])] ^ S6[IC(z[0])] ^ S7[IB(z[0])] ^ S8[ID(z[0])] ^ S8[IC(x[2])]; \ + z[2] = x[3] ^ S5[ID(z[1])] ^ S6[IC(z[1])] ^ S7[IB(z[1])] ^ S8[IA(z[1])] ^ S5[IB(x[2])]; \ + z[3] = x[1] ^ S5[IC(z[2])] ^ S6[IB(z[2])] ^ S7[ID(z[2])] ^ S8[IA(z[2])] ^ S6[ID(x[2])]; \ + } while (0) + +#define COMPUTE_X \ + do { \ + x[0] = z[2] ^ S5[IB(z[1])] ^ S6[ID(z[1])] ^ S7[IA(z[1])] ^ S8[IC(z[1])] ^ S7[IA(z[0])]; \ + x[1] = z[0] ^ S5[IA(x[0])] ^ S6[IC(x[0])] ^ S7[IB(x[0])] ^ S8[ID(x[0])] ^ S8[IC(z[0])]; \ + x[2] = z[1] ^ S5[ID(x[1])] ^ S6[IC(x[1])] ^ S7[IB(x[1])] ^ S8[IA(x[1])] ^ S5[IB(z[0])]; \ + x[3] = z[3] ^ S5[IC(x[2])] ^ S6[IB(x[2])] ^ S7[ID(x[2])] ^ S8[IA(x[2])] ^ S6[ID(z[0])]; \ + } while (0) + + +typedef struct AVCAST5 { + uint32_t Km[17]; + uint32_t Kr[17]; + int rounds; +} AVCAST5; + +const int av_cast5_size = sizeof(AVCAST5); + +static const uint32_t S1[256] = { + 0x30fb40d4, 0x9fa0ff0b, 0x6beccd2f, 0x3f258c7a, 0x1e213f2f, 0x9c004dd3, 0x6003e540, 0xcf9fc949, + 0xbfd4af27, 0x88bbbdb5, 0xe2034090, 0x98d09675, 0x6e63a0e0, 0x15c361d2, 0xc2e7661d, 0x22d4ff8e, + 0x28683b6f, 0xc07fd059, 0xff2379c8, 0x775f50e2, 0x43c340d3, 0xdf2f8656, 0x887ca41a, 0xa2d2bd2d, + 0xa1c9e0d6, 0x346c4819, 0x61b76d87, 0x22540f2f, 0x2abe32e1, 0xaa54166b, 0x22568e3a, 0xa2d341d0, + 0x66db40c8, 0xa784392f, 0x004dff2f, 0x2db9d2de, 0x97943fac, 0x4a97c1d8, 0x527644b7, 0xb5f437a7, + 0xb82cbaef, 0xd751d159, 0x6ff7f0ed, 0x5a097a1f, 0x827b68d0, 0x90ecf52e, 0x22b0c054, 0xbc8e5935, + 0x4b6d2f7f, 0x50bb64a2, 0xd2664910, 0xbee5812d, 0xb7332290, 0xe93b159f, 0xb48ee411, 0x4bff345d, + 0xfd45c240, 0xad31973f, 0xc4f6d02e, 0x55fc8165, 0xd5b1caad, 0xa1ac2dae, 0xa2d4b76d, 0xc19b0c50, + 0x882240f2, 0x0c6e4f38, 0xa4e4bfd7, 0x4f5ba272, 0x564c1d2f, 0xc59c5319, 0xb949e354, 0xb04669fe, + 0xb1b6ab8a, 0xc71358dd, 0x6385c545, 0x110f935d, 0x57538ad5, 0x6a390493, 0xe63d37e0, 0x2a54f6b3, + 0x3a787d5f, 0x6276a0b5, 0x19a6fcdf, 0x7a42206a, 0x29f9d4d5, 0xf61b1891, 0xbb72275e, 0xaa508167, + 0x38901091, 0xc6b505eb, 0x84c7cb8c, 0x2ad75a0f, 0x874a1427, 0xa2d1936b, 0x2ad286af, 0xaa56d291, + 0xd7894360, 0x425c750d, 0x93b39e26, 0x187184c9, 0x6c00b32d, 0x73e2bb14, 0xa0bebc3c, 0x54623779, + 0x64459eab, 0x3f328b82, 0x7718cf82, 0x59a2cea6, 0x04ee002e, 0x89fe78e6, 0x3fab0950, 0x325ff6c2, + 0x81383f05, 0x6963c5c8, 0x76cb5ad6, 0xd49974c9, 0xca180dcf, 0x380782d5, 0xc7fa5cf6, 0x8ac31511, + 0x35e79e13, 0x47da91d0, 0xf40f9086, 0xa7e2419e, 0x31366241, 0x051ef495, 0xaa573b04, 0x4a805d8d, + 0x548300d0, 0x00322a3c, 0xbf64cddf, 0xba57a68e, 0x75c6372b, 0x50afd341, 0xa7c13275, 0x915a0bf5, + 0x6b54bfab, 0x2b0b1426, 0xab4cc9d7, 0x449ccd82, 0xf7fbf265, 0xab85c5f3, 0x1b55db94, 0xaad4e324, + 0xcfa4bd3f, 0x2deaa3e2, 0x9e204d02, 0xc8bd25ac, 0xeadf55b3, 0xd5bd9e98, 0xe31231b2, 0x2ad5ad6c, + 0x954329de, 0xadbe4528, 0xd8710f69, 0xaa51c90f, 0xaa786bf6, 0x22513f1e, 0xaa51a79b, 0x2ad344cc, + 0x7b5a41f0, 0xd37cfbad, 0x1b069505, 0x41ece491, 0xb4c332e6, 0x032268d4, 0xc9600acc, 0xce387e6d, + 0xbf6bb16c, 0x6a70fb78, 0x0d03d9c9, 0xd4df39de, 0xe01063da, 0x4736f464, 0x5ad328d8, 0xb347cc96, + 0x75bb0fc3, 0x98511bfb, 0x4ffbcc35, 0xb58bcf6a, 0xe11f0abc, 0xbfc5fe4a, 0xa70aec10, 0xac39570a, + 0x3f04442f, 0x6188b153, 0xe0397a2e, 0x5727cb79, 0x9ceb418f, 0x1cacd68d, 0x2ad37c96, 0x0175cb9d, + 0xc69dff09, 0xc75b65f0, 0xd9db40d8, 0xec0e7779, 0x4744ead4, 0xb11c3274, 0xdd24cb9e, 0x7e1c54bd, + 0xf01144f9, 0xd2240eb1, 0x9675b3fd, 0xa3ac3755, 0xd47c27af, 0x51c85f4d, 0x56907596, 0xa5bb15e6, + 0x580304f0, 0xca042cf1, 0x011a37ea, 0x8dbfaadb, 0x35ba3e4a, 0x3526ffa0, 0xc37b4d09, 0xbc306ed9, + 0x98a52666, 0x5648f725, 0xff5e569d, 0x0ced63d0, 0x7c63b2cf, 0x700b45e1, 0xd5ea50f1, 0x85a92872, + 0xaf1fbda7, 0xd4234870, 0xa7870bf3, 0x2d3b4d79, 0x42e04198, 0x0cd0ede7, 0x26470db8, 0xf881814c, + 0x474d6ad7, 0x7c0c5e5c, 0xd1231959, 0x381b7298, 0xf5d2f4db, 0xab838653, 0x6e2f1e23, 0x83719c9e, + 0xbd91e046, 0x9a56456e, 0xdc39200c, 0x20c8c571, 0x962bda1c, 0xe1e696ff, 0xb141ab08, 0x7cca89b9, + 0x1a69e783, 0x02cc4843, 0xa2f7c579, 0x429ef47d, 0x427b169c, 0x5ac9f049, 0xdd8f0f00, 0x5c8165bf +}; + +static const uint32_t S2[256] = { + 0x1f201094, 0xef0ba75b, 0x69e3cf7e, 0x393f4380, 0xfe61cf7a, 0xeec5207a, 0x55889c94, 0x72fc0651, + 0xada7ef79, 0x4e1d7235, 0xd55a63ce, 0xde0436ba, 0x99c430ef, 0x5f0c0794, 0x18dcdb7d, 0xa1d6eff3, + 0xa0b52f7b, 0x59e83605, 0xee15b094, 0xe9ffd909, 0xdc440086, 0xef944459, 0xba83ccb3, 0xe0c3cdfb, + 0xd1da4181, 0x3b092ab1, 0xf997f1c1, 0xa5e6cf7b, 0x01420ddb, 0xe4e7ef5b, 0x25a1ff41, 0xe180f806, + 0x1fc41080, 0x179bee7a, 0xd37ac6a9, 0xfe5830a4, 0x98de8b7f, 0x77e83f4e, 0x79929269, 0x24fa9f7b, + 0xe113c85b, 0xacc40083, 0xd7503525, 0xf7ea615f, 0x62143154, 0x0d554b63, 0x5d681121, 0xc866c359, + 0x3d63cf73, 0xcee234c0, 0xd4d87e87, 0x5c672b21, 0x071f6181, 0x39f7627f, 0x361e3084, 0xe4eb573b, + 0x602f64a4, 0xd63acd9c, 0x1bbc4635, 0x9e81032d, 0x2701f50c, 0x99847ab4, 0xa0e3df79, 0xba6cf38c, + 0x10843094, 0x2537a95e, 0xf46f6ffe, 0xa1ff3b1f, 0x208cfb6a, 0x8f458c74, 0xd9e0a227, 0x4ec73a34, + 0xfc884f69, 0x3e4de8df, 0xef0e0088, 0x3559648d, 0x8a45388c, 0x1d804366, 0x721d9bfd, 0xa58684bb, + 0xe8256333, 0x844e8212, 0x128d8098, 0xfed33fb4, 0xce280ae1, 0x27e19ba5, 0xd5a6c252, 0xe49754bd, + 0xc5d655dd, 0xeb667064, 0x77840b4d, 0xa1b6a801, 0x84db26a9, 0xe0b56714, 0x21f043b7, 0xe5d05860, + 0x54f03084, 0x066ff472, 0xa31aa153, 0xdadc4755, 0xb5625dbf, 0x68561be6, 0x83ca6b94, 0x2d6ed23b, + 0xeccf01db, 0xa6d3d0ba, 0xb6803d5c, 0xaf77a709, 0x33b4a34c, 0x397bc8d6, 0x5ee22b95, 0x5f0e5304, + 0x81ed6f61, 0x20e74364, 0xb45e1378, 0xde18639b, 0x881ca122, 0xb96726d1, 0x8049a7e8, 0x22b7da7b, + 0x5e552d25, 0x5272d237, 0x79d2951c, 0xc60d894c, 0x488cb402, 0x1ba4fe5b, 0xa4b09f6b, 0x1ca815cf, + 0xa20c3005, 0x8871df63, 0xb9de2fcb, 0x0cc6c9e9, 0x0beeff53, 0xe3214517, 0xb4542835, 0x9f63293c, + 0xee41e729, 0x6e1d2d7c, 0x50045286, 0x1e6685f3, 0xf33401c6, 0x30a22c95, 0x31a70850, 0x60930f13, + 0x73f98417, 0xa1269859, 0xec645c44, 0x52c877a9, 0xcdff33a6, 0xa02b1741, 0x7cbad9a2, 0x2180036f, + 0x50d99c08, 0xcb3f4861, 0xc26bd765, 0x64a3f6ab, 0x80342676, 0x25a75e7b, 0xe4e6d1fc, 0x20c710e6, + 0xcdf0b680, 0x17844d3b, 0x31eef84d, 0x7e0824e4, 0x2ccb49eb, 0x846a3bae, 0x8ff77888, 0xee5d60f6, + 0x7af75673, 0x2fdd5cdb, 0xa11631c1, 0x30f66f43, 0xb3faec54, 0x157fd7fa, 0xef8579cc, 0xd152de58, + 0xdb2ffd5e, 0x8f32ce19, 0x306af97a, 0x02f03ef8, 0x99319ad5, 0xc242fa0f, 0xa7e3ebb0, 0xc68e4906, + 0xb8da230c, 0x80823028, 0xdcdef3c8, 0xd35fb171, 0x088a1bc8, 0xbec0c560, 0x61a3c9e8, 0xbca8f54d, + 0xc72feffa, 0x22822e99, 0x82c570b4, 0xd8d94e89, 0x8b1c34bc, 0x301e16e6, 0x273be979, 0xb0ffeaa6, + 0x61d9b8c6, 0x00b24869, 0xb7ffce3f, 0x08dc283b, 0x43daf65a, 0xf7e19798, 0x7619b72f, 0x8f1c9ba4, + 0xdc8637a0, 0x16a7d3b1, 0x9fc393b7, 0xa7136eeb, 0xc6bcc63e, 0x1a513742, 0xef6828bc, 0x520365d6, + 0x2d6a77ab, 0x3527ed4b, 0x821fd216, 0x095c6e2e, 0xdb92f2fb, 0x5eea29cb, 0x145892f5, 0x91584f7f, + 0x5483697b, 0x2667a8cc, 0x85196048, 0x8c4bacea, 0x833860d4, 0x0d23e0f9, 0x6c387e8a, 0x0ae6d249, + 0xb284600c, 0xd835731d, 0xdcb1c647, 0xac4c56ea, 0x3ebd81b3, 0x230eabb0, 0x6438bc87, 0xf0b5b1fa, + 0x8f5ea2b3, 0xfc184642, 0x0a036b7a, 0x4fb089bd, 0x649da589, 0xa345415e, 0x5c038323, 0x3e5d3bb9, + 0x43d79572, 0x7e6dd07c, 0x06dfdf1e, 0x6c6cc4ef, 0x7160a539, 0x73bfbe70, 0x83877605, 0x4523ecf1 +}; + +static const uint32_t S3[256] = { + 0x8defc240, 0x25fa5d9f, 0xeb903dbf, 0xe810c907, 0x47607fff, 0x369fe44b, 0x8c1fc644, 0xaececa90, + 0xbeb1f9bf, 0xeefbcaea, 0xe8cf1950, 0x51df07ae, 0x920e8806, 0xf0ad0548, 0xe13c8d83, 0x927010d5, + 0x11107d9f, 0x07647db9, 0xb2e3e4d4, 0x3d4f285e, 0xb9afa820, 0xfade82e0, 0xa067268b, 0x8272792e, + 0x553fb2c0, 0x489ae22b, 0xd4ef9794, 0x125e3fbc, 0x21fffcee, 0x825b1bfd, 0x9255c5ed, 0x1257a240, + 0x4e1a8302, 0xbae07fff, 0x528246e7, 0x8e57140e, 0x3373f7bf, 0x8c9f8188, 0xa6fc4ee8, 0xc982b5a5, + 0xa8c01db7, 0x579fc264, 0x67094f31, 0xf2bd3f5f, 0x40fff7c1, 0x1fb78dfc, 0x8e6bd2c1, 0x437be59b, + 0x99b03dbf, 0xb5dbc64b, 0x638dc0e6, 0x55819d99, 0xa197c81c, 0x4a012d6e, 0xc5884a28, 0xccc36f71, + 0xb843c213, 0x6c0743f1, 0x8309893c, 0x0feddd5f, 0x2f7fe850, 0xd7c07f7e, 0x02507fbf, 0x5afb9a04, + 0xa747d2d0, 0x1651192e, 0xaf70bf3e, 0x58c31380, 0x5f98302e, 0x727cc3c4, 0x0a0fb402, 0x0f7fef82, + 0x8c96fdad, 0x5d2c2aae, 0x8ee99a49, 0x50da88b8, 0x8427f4a0, 0x1eac5790, 0x796fb449, 0x8252dc15, + 0xefbd7d9b, 0xa672597d, 0xada840d8, 0x45f54504, 0xfa5d7403, 0xe83ec305, 0x4f91751a, 0x925669c2, + 0x23efe941, 0xa903f12e, 0x60270df2, 0x0276e4b6, 0x94fd6574, 0x927985b2, 0x8276dbcb, 0x02778176, + 0xf8af918d, 0x4e48f79e, 0x8f616ddf, 0xe29d840e, 0x842f7d83, 0x340ce5c8, 0x96bbb682, 0x93b4b148, + 0xef303cab, 0x984faf28, 0x779faf9b, 0x92dc560d, 0x224d1e20, 0x8437aa88, 0x7d29dc96, 0x2756d3dc, + 0x8b907cee, 0xb51fd240, 0xe7c07ce3, 0xe566b4a1, 0xc3e9615e, 0x3cf8209d, 0x6094d1e3, 0xcd9ca341, + 0x5c76460e, 0x00ea983b, 0xd4d67881, 0xfd47572c, 0xf76cedd9, 0xbda8229c, 0x127dadaa, 0x438a074e, + 0x1f97c090, 0x081bdb8a, 0x93a07ebe, 0xb938ca15, 0x97b03cff, 0x3dc2c0f8, 0x8d1ab2ec, 0x64380e51, + 0x68cc7bfb, 0xd90f2788, 0x12490181, 0x5de5ffd4, 0xdd7ef86a, 0x76a2e214, 0xb9a40368, 0x925d958f, + 0x4b39fffa, 0xba39aee9, 0xa4ffd30b, 0xfaf7933b, 0x6d498623, 0x193cbcfa, 0x27627545, 0x825cf47a, + 0x61bd8ba0, 0xd11e42d1, 0xcead04f4, 0x127ea392, 0x10428db7, 0x8272a972, 0x9270c4a8, 0x127de50b, + 0x285ba1c8, 0x3c62f44f, 0x35c0eaa5, 0xe805d231, 0x428929fb, 0xb4fcdf82, 0x4fb66a53, 0x0e7dc15b, + 0x1f081fab, 0x108618ae, 0xfcfd086d, 0xf9ff2889, 0x694bcc11, 0x236a5cae, 0x12deca4d, 0x2c3f8cc5, + 0xd2d02dfe, 0xf8ef5896, 0xe4cf52da, 0x95155b67, 0x494a488c, 0xb9b6a80c, 0x5c8f82bc, 0x89d36b45, + 0x3a609437, 0xec00c9a9, 0x44715253, 0x0a874b49, 0xd773bc40, 0x7c34671c, 0x02717ef6, 0x4feb5536, + 0xa2d02fff, 0xd2bf60c4, 0xd43f03c0, 0x50b4ef6d, 0x07478cd1, 0x006e1888, 0xa2e53f55, 0xb9e6d4bc, + 0xa2048016, 0x97573833, 0xd7207d67, 0xde0f8f3d, 0x72f87b33, 0xabcc4f33, 0x7688c55d, 0x7b00a6b0, + 0x947b0001, 0x570075d2, 0xf9bb88f8, 0x8942019e, 0x4264a5ff, 0x856302e0, 0x72dbd92b, 0xee971b69, + 0x6ea22fde, 0x5f08ae2b, 0xaf7a616d, 0xe5c98767, 0xcf1febd2, 0x61efc8c2, 0xf1ac2571, 0xcc8239c2, + 0x67214cb8, 0xb1e583d1, 0xb7dc3e62, 0x7f10bdce, 0xf90a5c38, 0x0ff0443d, 0x606e6dc6, 0x60543a49, + 0x5727c148, 0x2be98a1d, 0x8ab41738, 0x20e1be24, 0xaf96da0f, 0x68458425, 0x99833be5, 0x600d457d, + 0x282f9350, 0x8334b362, 0xd91d1120, 0x2b6d8da0, 0x642b1e31, 0x9c305a00, 0x52bce688, 0x1b03588a, + 0xf7baefd5, 0x4142ed9c, 0xa4315c11, 0x83323ec5, 0xdfef4636, 0xa133c501, 0xe9d3531c, 0xee353783 +}; + +static const uint32_t S4[256] = { + 0x9db30420, 0x1fb6e9de, 0xa7be7bef, 0xd273a298, 0x4a4f7bdb, 0x64ad8c57, 0x85510443, 0xfa020ed1, + 0x7e287aff, 0xe60fb663, 0x095f35a1, 0x79ebf120, 0xfd059d43, 0x6497b7b1, 0xf3641f63, 0x241e4adf, + 0x28147f5f, 0x4fa2b8cd, 0xc9430040, 0x0cc32220, 0xfdd30b30, 0xc0a5374f, 0x1d2d00d9, 0x24147b15, + 0xee4d111a, 0x0fca5167, 0x71ff904c, 0x2d195ffe, 0x1a05645f, 0x0c13fefe, 0x081b08ca, 0x05170121, + 0x80530100, 0xe83e5efe, 0xac9af4f8, 0x7fe72701, 0xd2b8ee5f, 0x06df4261, 0xbb9e9b8a, 0x7293ea25, + 0xce84ffdf, 0xf5718801, 0x3dd64b04, 0xa26f263b, 0x7ed48400, 0x547eebe6, 0x446d4ca0, 0x6cf3d6f5, + 0x2649abdf, 0xaea0c7f5, 0x36338cc1, 0x503f7e93, 0xd3772061, 0x11b638e1, 0x72500e03, 0xf80eb2bb, + 0xabe0502e, 0xec8d77de, 0x57971e81, 0xe14f6746, 0xc9335400, 0x6920318f, 0x081dbb99, 0xffc304a5, + 0x4d351805, 0x7f3d5ce3, 0xa6c866c6, 0x5d5bcca9, 0xdaec6fea, 0x9f926f91, 0x9f46222f, 0x3991467d, + 0xa5bf6d8e, 0x1143c44f, 0x43958302, 0xd0214eeb, 0x022083b8, 0x3fb6180c, 0x18f8931e, 0x281658e6, + 0x26486e3e, 0x8bd78a70, 0x7477e4c1, 0xb506e07c, 0xf32d0a25, 0x79098b02, 0xe4eabb81, 0x28123b23, + 0x69dead38, 0x1574ca16, 0xdf871b62, 0x211c40b7, 0xa51a9ef9, 0x0014377b, 0x041e8ac8, 0x09114003, + 0xbd59e4d2, 0xe3d156d5, 0x4fe876d5, 0x2f91a340, 0x557be8de, 0x00eae4a7, 0x0ce5c2ec, 0x4db4bba6, + 0xe756bdff, 0xdd3369ac, 0xec17b035, 0x06572327, 0x99afc8b0, 0x56c8c391, 0x6b65811c, 0x5e146119, + 0x6e85cb75, 0xbe07c002, 0xc2325577, 0x893ff4ec, 0x5bbfc92d, 0xd0ec3b25, 0xb7801ab7, 0x8d6d3b24, + 0x20c763ef, 0xc366a5fc, 0x9c382880, 0x0ace3205, 0xaac9548a, 0xeca1d7c7, 0x041afa32, 0x1d16625a, + 0x6701902c, 0x9b757a54, 0x31d477f7, 0x9126b031, 0x36cc6fdb, 0xc70b8b46, 0xd9e66a48, 0x56e55a79, + 0x026a4ceb, 0x52437eff, 0x2f8f76b4, 0x0df980a5, 0x8674cde3, 0xedda04eb, 0x17a9be04, 0x2c18f4df, + 0xb7747f9d, 0xab2af7b4, 0xefc34d20, 0x2e096b7c, 0x1741a254, 0xe5b6a035, 0x213d42f6, 0x2c1c7c26, + 0x61c2f50f, 0x6552daf9, 0xd2c231f8, 0x25130f69, 0xd8167fa2, 0x0418f2c8, 0x001a96a6, 0x0d1526ab, + 0x63315c21, 0x5e0a72ec, 0x49bafefd, 0x187908d9, 0x8d0dbd86, 0x311170a7, 0x3e9b640c, 0xcc3e10d7, + 0xd5cad3b6, 0x0caec388, 0xf73001e1, 0x6c728aff, 0x71eae2a1, 0x1f9af36e, 0xcfcbd12f, 0xc1de8417, + 0xac07be6b, 0xcb44a1d8, 0x8b9b0f56, 0x013988c3, 0xb1c52fca, 0xb4be31cd, 0xd8782806, 0x12a3a4e2, + 0x6f7de532, 0x58fd7eb6, 0xd01ee900, 0x24adffc2, 0xf4990fc5, 0x9711aac5, 0x001d7b95, 0x82e5e7d2, + 0x109873f6, 0x00613096, 0xc32d9521, 0xada121ff, 0x29908415, 0x7fbb977f, 0xaf9eb3db, 0x29c9ed2a, + 0x5ce2a465, 0xa730f32c, 0xd0aa3fe8, 0x8a5cc091, 0xd49e2ce7, 0x0ce454a9, 0xd60acd86, 0x015f1919, + 0x77079103, 0xdea03af6, 0x78a8565e, 0xdee356df, 0x21f05cbe, 0x8b75e387, 0xb3c50651, 0xb8a5c3ef, + 0xd8eeb6d2, 0xe523be77, 0xc2154529, 0x2f69efdf, 0xafe67afb, 0xf470c4b2, 0xf3e0eb5b, 0xd6cc9876, + 0x39e4460c, 0x1fda8538, 0x1987832f, 0xca007367, 0xa99144f8, 0x296b299e, 0x492fc295, 0x9266beab, + 0xb5676e69, 0x9bd3ddda, 0xdf7e052f, 0xdb25701c, 0x1b5e51ee, 0xf65324e6, 0x6afce36c, 0x0316cc04, + 0x8644213e, 0xb7dc59d0, 0x7965291f, 0xccd6fd43, 0x41823979, 0x932bcdf6, 0xb657c34d, 0x4edfd282, + 0x7ae5290c, 0x3cb9536b, 0x851e20fe, 0x9833557e, 0x13ecf0b0, 0xd3ffb372, 0x3f85c5c1, 0x0aef7ed2 +}; + +static const uint32_t S5[256] = { + 0x7ec90c04, 0x2c6e74b9, 0x9b0e66df, 0xa6337911, 0xb86a7fff, 0x1dd358f5, 0x44dd9d44, 0x1731167f, + 0x08fbf1fa, 0xe7f511cc, 0xd2051b00, 0x735aba00, 0x2ab722d8, 0x386381cb, 0xacf6243a, 0x69befd7a, + 0xe6a2e77f, 0xf0c720cd, 0xc4494816, 0xccf5c180, 0x38851640, 0x15b0a848, 0xe68b18cb, 0x4caadeff, + 0x5f480a01, 0x0412b2aa, 0x259814fc, 0x41d0efe2, 0x4e40b48d, 0x248eb6fb, 0x8dba1cfe, 0x41a99b02, + 0x1a550a04, 0xba8f65cb, 0x7251f4e7, 0x95a51725, 0xc106ecd7, 0x97a5980a, 0xc539b9aa, 0x4d79fe6a, + 0xf2f3f763, 0x68af8040, 0xed0c9e56, 0x11b4958b, 0xe1eb5a88, 0x8709e6b0, 0xd7e07156, 0x4e29fea7, + 0x6366e52d, 0x02d1c000, 0xc4ac8e05, 0x9377f571, 0x0c05372a, 0x578535f2, 0x2261be02, 0xd642a0c9, + 0xdf13a280, 0x74b55bd2, 0x682199c0, 0xd421e5ec, 0x53fb3ce8, 0xc8adedb3, 0x28a87fc9, 0x3d959981, + 0x5c1ff900, 0xfe38d399, 0x0c4eff0b, 0x062407ea, 0xaa2f4fb1, 0x4fb96976, 0x90c79505, 0xb0a8a774, + 0xef55a1ff, 0xe59ca2c2, 0xa6b62d27, 0xe66a4263, 0xdf65001f, 0x0ec50966, 0xdfdd55bc, 0x29de0655, + 0x911e739a, 0x17af8975, 0x32c7911c, 0x89f89468, 0x0d01e980, 0x524755f4, 0x03b63cc9, 0x0cc844b2, + 0xbcf3f0aa, 0x87ac36e9, 0xe53a7426, 0x01b3d82b, 0x1a9e7449, 0x64ee2d7e, 0xcddbb1da, 0x01c94910, + 0xb868bf80, 0x0d26f3fd, 0x9342ede7, 0x04a5c284, 0x636737b6, 0x50f5b616, 0xf24766e3, 0x8eca36c1, + 0x136e05db, 0xfef18391, 0xfb887a37, 0xd6e7f7d4, 0xc7fb7dc9, 0x3063fcdf, 0xb6f589de, 0xec2941da, + 0x26e46695, 0xb7566419, 0xf654efc5, 0xd08d58b7, 0x48925401, 0xc1bacb7f, 0xe5ff550f, 0xb6083049, + 0x5bb5d0e8, 0x87d72e5a, 0xab6a6ee1, 0x223a66ce, 0xc62bf3cd, 0x9e0885f9, 0x68cb3e47, 0x086c010f, + 0xa21de820, 0xd18b69de, 0xf3f65777, 0xfa02c3f6, 0x407edac3, 0xcbb3d550, 0x1793084d, 0xb0d70eba, + 0x0ab378d5, 0xd951fb0c, 0xded7da56, 0x4124bbe4, 0x94ca0b56, 0x0f5755d1, 0xe0e1e56e, 0x6184b5be, + 0x580a249f, 0x94f74bc0, 0xe327888e, 0x9f7b5561, 0xc3dc0280, 0x05687715, 0x646c6bd7, 0x44904db3, + 0x66b4f0a3, 0xc0f1648a, 0x697ed5af, 0x49e92ff6, 0x309e374f, 0x2cb6356a, 0x85808573, 0x4991f840, + 0x76f0ae02, 0x083be84d, 0x28421c9a, 0x44489406, 0x736e4cb8, 0xc1092910, 0x8bc95fc6, 0x7d869cf4, + 0x134f616f, 0x2e77118d, 0xb31b2be1, 0xaa90b472, 0x3ca5d717, 0x7d161bba, 0x9cad9010, 0xaf462ba2, + 0x9fe459d2, 0x45d34559, 0xd9f2da13, 0xdbc65487, 0xf3e4f94e, 0x176d486f, 0x097c13ea, 0x631da5c7, + 0x445f7382, 0x175683f4, 0xcdc66a97, 0x70be0288, 0xb3cdcf72, 0x6e5dd2f3, 0x20936079, 0x459b80a5, + 0xbe60e2db, 0xa9c23101, 0xeba5315c, 0x224e42f2, 0x1c5c1572, 0xf6721b2c, 0x1ad2fff3, 0x8c25404e, + 0x324ed72f, 0x4067b7fd, 0x0523138e, 0x5ca3bc78, 0xdc0fd66e, 0x75922283, 0x784d6b17, 0x58ebb16e, + 0x44094f85, 0x3f481d87, 0xfcfeae7b, 0x77b5ff76, 0x8c2302bf, 0xaaf47556, 0x5f46b02a, 0x2b092801, + 0x3d38f5f7, 0x0ca81f36, 0x52af4a8a, 0x66d5e7c0, 0xdf3b0874, 0x95055110, 0x1b5ad7a8, 0xf61ed5ad, + 0x6cf6e479, 0x20758184, 0xd0cefa65, 0x88f7be58, 0x4a046826, 0x0ff6f8f3, 0xa09c7f70, 0x5346aba0, + 0x5ce96c28, 0xe176eda3, 0x6bac307f, 0x376829d2, 0x85360fa9, 0x17e3fe2a, 0x24b79767, 0xf5a96b20, + 0xd6cd2595, 0x68ff1ebf, 0x7555442c, 0xf19f06be, 0xf9e0659a, 0xeeb9491d, 0x34010718, 0xbb30cab8, + 0xe822fe15, 0x88570983, 0x750e6249, 0xda627e55, 0x5e76ffa8, 0xb1534546, 0x6d47de08, 0xefe9e7d4 +}; + +static const uint32_t S6[256] = { + 0xf6fa8f9d, 0x2cac6ce1, 0x4ca34867, 0xe2337f7c, 0x95db08e7, 0x016843b4, 0xeced5cbc, 0x325553ac, + 0xbf9f0960, 0xdfa1e2ed, 0x83f0579d, 0x63ed86b9, 0x1ab6a6b8, 0xde5ebe39, 0xf38ff732, 0x8989b138, + 0x33f14961, 0xc01937bd, 0xf506c6da, 0xe4625e7e, 0xa308ea99, 0x4e23e33c, 0x79cbd7cc, 0x48a14367, + 0xa3149619, 0xfec94bd5, 0xa114174a, 0xeaa01866, 0xa084db2d, 0x09a8486f, 0xa888614a, 0x2900af98, + 0x01665991, 0xe1992863, 0xc8f30c60, 0x2e78ef3c, 0xd0d51932, 0xcf0fec14, 0xf7ca07d2, 0xd0a82072, + 0xfd41197e, 0x9305a6b0, 0xe86be3da, 0x74bed3cd, 0x372da53c, 0x4c7f4448, 0xdab5d440, 0x6dba0ec3, + 0x083919a7, 0x9fbaeed9, 0x49dbcfb0, 0x4e670c53, 0x5c3d9c01, 0x64bdb941, 0x2c0e636a, 0xba7dd9cd, + 0xea6f7388, 0xe70bc762, 0x35f29adb, 0x5c4cdd8d, 0xf0d48d8c, 0xb88153e2, 0x08a19866, 0x1ae2eac8, + 0x284caf89, 0xaa928223, 0x9334be53, 0x3b3a21bf, 0x16434be3, 0x9aea3906, 0xefe8c36e, 0xf890cdd9, + 0x80226dae, 0xc340a4a3, 0xdf7e9c09, 0xa694a807, 0x5b7c5ecc, 0x221db3a6, 0x9a69a02f, 0x68818a54, + 0xceb2296f, 0x53c0843a, 0xfe893655, 0x25bfe68a, 0xb4628abc, 0xcf222ebf, 0x25ac6f48, 0xa9a99387, + 0x53bddb65, 0xe76ffbe7, 0xe967fd78, 0x0ba93563, 0x8e342bc1, 0xe8a11be9, 0x4980740d, 0xc8087dfc, + 0x8de4bf99, 0xa11101a0, 0x7fd37975, 0xda5a26c0, 0xe81f994f, 0x9528cd89, 0xfd339fed, 0xb87834bf, + 0x5f04456d, 0x22258698, 0xc9c4c83b, 0x2dc156be, 0x4f628daa, 0x57f55ec5, 0xe2220abe, 0xd2916ebf, + 0x4ec75b95, 0x24f2c3c0, 0x42d15d99, 0xcd0d7fa0, 0x7b6e27ff, 0xa8dc8af0, 0x7345c106, 0xf41e232f, + 0x35162386, 0xe6ea8926, 0x3333b094, 0x157ec6f2, 0x372b74af, 0x692573e4, 0xe9a9d848, 0xf3160289, + 0x3a62ef1d, 0xa787e238, 0xf3a5f676, 0x74364853, 0x20951063, 0x4576698d, 0xb6fad407, 0x592af950, + 0x36f73523, 0x4cfb6e87, 0x7da4cec0, 0x6c152daa, 0xcb0396a8, 0xc50dfe5d, 0xfcd707ab, 0x0921c42f, + 0x89dff0bb, 0x5fe2be78, 0x448f4f33, 0x754613c9, 0x2b05d08d, 0x48b9d585, 0xdc049441, 0xc8098f9b, + 0x7dede786, 0xc39a3373, 0x42410005, 0x6a091751, 0x0ef3c8a6, 0x890072d6, 0x28207682, 0xa9a9f7be, + 0xbf32679d, 0xd45b5b75, 0xb353fd00, 0xcbb0e358, 0x830f220a, 0x1f8fb214, 0xd372cf08, 0xcc3c4a13, + 0x8cf63166, 0x061c87be, 0x88c98f88, 0x6062e397, 0x47cf8e7a, 0xb6c85283, 0x3cc2acfb, 0x3fc06976, + 0x4e8f0252, 0x64d8314d, 0xda3870e3, 0x1e665459, 0xc10908f0, 0x513021a5, 0x6c5b68b7, 0x822f8aa0, + 0x3007cd3e, 0x74719eef, 0xdc872681, 0x073340d4, 0x7e432fd9, 0x0c5ec241, 0x8809286c, 0xf592d891, + 0x08a930f6, 0x957ef305, 0xb7fbffbd, 0xc266e96f, 0x6fe4ac98, 0xb173ecc0, 0xbc60b42a, 0x953498da, + 0xfba1ae12, 0x2d4bd736, 0x0f25faab, 0xa4f3fceb, 0xe2969123, 0x257f0c3d, 0x9348af49, 0x361400bc, + 0xe8816f4a, 0x3814f200, 0xa3f94043, 0x9c7a54c2, 0xbc704f57, 0xda41e7f9, 0xc25ad33a, 0x54f4a084, + 0xb17f5505, 0x59357cbe, 0xedbd15c8, 0x7f97c5ab, 0xba5ac7b5, 0xb6f6deaf, 0x3a479c3a, 0x5302da25, + 0x653d7e6a, 0x54268d49, 0x51a477ea, 0x5017d55b, 0xd7d25d88, 0x44136c76, 0x0404a8c8, 0xb8e5a121, + 0xb81a928a, 0x60ed5869, 0x97c55b96, 0xeaec991b, 0x29935913, 0x01fdb7f1, 0x088e8dfa, 0x9ab6f6f5, + 0x3b4cbf9f, 0x4a5de3ab, 0xe6051d35, 0xa0e1d855, 0xd36b4cf1, 0xf544edeb, 0xb0e93524, 0xbebb8fbd, + 0xa2d762cf, 0x49c92f54, 0x38b5f331, 0x7128a454, 0x48392905, 0xa65b1db8, 0x851c97bd, 0xd675cf2f +}; + +static const uint32_t S7[256] = { + 0x85e04019, 0x332bf567, 0x662dbfff, 0xcfc65693, 0x2a8d7f6f, 0xab9bc912, 0xde6008a1, 0x2028da1f, + 0x0227bce7, 0x4d642916, 0x18fac300, 0x50f18b82, 0x2cb2cb11, 0xb232e75c, 0x4b3695f2, 0xb28707de, + 0xa05fbcf6, 0xcd4181e9, 0xe150210c, 0xe24ef1bd, 0xb168c381, 0xfde4e789, 0x5c79b0d8, 0x1e8bfd43, + 0x4d495001, 0x38be4341, 0x913cee1d, 0x92a79c3f, 0x089766be, 0xbaeeadf4, 0x1286becf, 0xb6eacb19, + 0x2660c200, 0x7565bde4, 0x64241f7a, 0x8248dca9, 0xc3b3ad66, 0x28136086, 0x0bd8dfa8, 0x356d1cf2, + 0x107789be, 0xb3b2e9ce, 0x0502aa8f, 0x0bc0351e, 0x166bf52a, 0xeb12ff82, 0xe3486911, 0xd34d7516, + 0x4e7b3aff, 0x5f43671b, 0x9cf6e037, 0x4981ac83, 0x334266ce, 0x8c9341b7, 0xd0d854c0, 0xcb3a6c88, + 0x47bc2829, 0x4725ba37, 0xa66ad22b, 0x7ad61f1e, 0x0c5cbafa, 0x4437f107, 0xb6e79962, 0x42d2d816, + 0x0a961288, 0xe1a5c06e, 0x13749e67, 0x72fc081a, 0xb1d139f7, 0xf9583745, 0xcf19df58, 0xbec3f756, + 0xc06eba30, 0x07211b24, 0x45c28829, 0xc95e317f, 0xbc8ec511, 0x38bc46e9, 0xc6e6fa14, 0xbae8584a, + 0xad4ebc46, 0x468f508b, 0x7829435f, 0xf124183b, 0x821dba9f, 0xaff60ff4, 0xea2c4e6d, 0x16e39264, + 0x92544a8b, 0x009b4fc3, 0xaba68ced, 0x9ac96f78, 0x06a5b79a, 0xb2856e6e, 0x1aec3ca9, 0xbe838688, + 0x0e0804e9, 0x55f1be56, 0xe7e5363b, 0xb3a1f25d, 0xf7debb85, 0x61fe033c, 0x16746233, 0x3c034c28, + 0xda6d0c74, 0x79aac56c, 0x3ce4e1ad, 0x51f0c802, 0x98f8f35a, 0x1626a49f, 0xeed82b29, 0x1d382fe3, + 0x0c4fb99a, 0xbb325778, 0x3ec6d97b, 0x6e77a6a9, 0xcb658b5c, 0xd45230c7, 0x2bd1408b, 0x60c03eb7, + 0xb9068d78, 0xa33754f4, 0xf430c87d, 0xc8a71302, 0xb96d8c32, 0xebd4e7be, 0xbe8b9d2d, 0x7979fb06, + 0xe7225308, 0x8b75cf77, 0x11ef8da4, 0xe083c858, 0x8d6b786f, 0x5a6317a6, 0xfa5cf7a0, 0x5dda0033, + 0xf28ebfb0, 0xf5b9c310, 0xa0eac280, 0x08b9767a, 0xa3d9d2b0, 0x79d34217, 0x021a718d, 0x9ac6336a, + 0x2711fd60, 0x438050e3, 0x069908a8, 0x3d7fedc4, 0x826d2bef, 0x4eeb8476, 0x488dcf25, 0x36c9d566, + 0x28e74e41, 0xc2610aca, 0x3d49a9cf, 0xbae3b9df, 0xb65f8de6, 0x92aeaf64, 0x3ac7d5e6, 0x9ea80509, + 0xf22b017d, 0xa4173f70, 0xdd1e16c3, 0x15e0d7f9, 0x50b1b887, 0x2b9f4fd5, 0x625aba82, 0x6a017962, + 0x2ec01b9c, 0x15488aa9, 0xd716e740, 0x40055a2c, 0x93d29a22, 0xe32dbf9a, 0x058745b9, 0x3453dc1e, + 0xd699296e, 0x496cff6f, 0x1c9f4986, 0xdfe2ed07, 0xb87242d1, 0x19de7eae, 0x053e561a, 0x15ad6f8c, + 0x66626c1c, 0x7154c24c, 0xea082b2a, 0x93eb2939, 0x17dcb0f0, 0x58d4f2ae, 0x9ea294fb, 0x52cf564c, + 0x9883fe66, 0x2ec40581, 0x763953c3, 0x01d6692e, 0xd3a0c108, 0xa1e7160e, 0xe4f2dfa6, 0x693ed285, + 0x74904698, 0x4c2b0edd, 0x4f757656, 0x5d393378, 0xa132234f, 0x3d321c5d, 0xc3f5e194, 0x4b269301, + 0xc79f022f, 0x3c997e7e, 0x5e4f9504, 0x3ffafbbd, 0x76f7ad0e, 0x296693f4, 0x3d1fce6f, 0xc61e45be, + 0xd3b5ab34, 0xf72bf9b7, 0x1b0434c0, 0x4e72b567, 0x5592a33d, 0xb5229301, 0xcfd2a87f, 0x60aeb767, + 0x1814386b, 0x30bcc33d, 0x38a0c07d, 0xfd1606f2, 0xc363519b, 0x589dd390, 0x5479f8e6, 0x1cb8d647, + 0x97fd61a9, 0xea7759f4, 0x2d57539d, 0x569a58cf, 0xe84e63ad, 0x462e1b78, 0x6580f87e, 0xf3817914, + 0x91da55f4, 0x40a230f3, 0xd1988f35, 0xb6e318d2, 0x3ffa50bc, 0x3d40f021, 0xc3c0bdae, 0x4958c24c, + 0x518f36b2, 0x84b1d370, 0x0fedce83, 0x878ddada, 0xf2a279c7, 0x94e01be8, 0x90716f4b, 0x954b8aa3 +}; + +static const uint32_t S8[256] = { + 0xe216300d, 0xbbddfffc, 0xa7ebdabd, 0x35648095, 0x7789f8b7, 0xe6c1121b, 0x0e241600, 0x052ce8b5, + 0x11a9cfb0, 0xe5952f11, 0xece7990a, 0x9386d174, 0x2a42931c, 0x76e38111, 0xb12def3a, 0x37ddddfc, + 0xde9adeb1, 0x0a0cc32c, 0xbe197029, 0x84a00940, 0xbb243a0f, 0xb4d137cf, 0xb44e79f0, 0x049eedfd, + 0x0b15a15d, 0x480d3168, 0x8bbbde5a, 0x669ded42, 0xc7ece831, 0x3f8f95e7, 0x72df191b, 0x7580330d, + 0x94074251, 0x5c7dcdfa, 0xabbe6d63, 0xaa402164, 0xb301d40a, 0x02e7d1ca, 0x53571dae, 0x7a3182a2, + 0x12a8ddec, 0xfdaa335d, 0x176f43e8, 0x71fb46d4, 0x38129022, 0xce949ad4, 0xb84769ad, 0x965bd862, + 0x82f3d055, 0x66fb9767, 0x15b80b4e, 0x1d5b47a0, 0x4cfde06f, 0xc28ec4b8, 0x57e8726e, 0x647a78fc, + 0x99865d44, 0x608bd593, 0x6c200e03, 0x39dc5ff6, 0x5d0b00a3, 0xae63aff2, 0x7e8bd632, 0x70108c0c, + 0xbbd35049, 0x2998df04, 0x980cf42a, 0x9b6df491, 0x9e7edd53, 0x06918548, 0x58cb7e07, 0x3b74ef2e, + 0x522fffb1, 0xd24708cc, 0x1c7e27cd, 0xa4eb215b, 0x3cf1d2e2, 0x19b47a38, 0x424f7618, 0x35856039, + 0x9d17dee7, 0x27eb35e6, 0xc9aff67b, 0x36baf5b8, 0x09c467cd, 0xc18910b1, 0xe11dbf7b, 0x06cd1af8, + 0x7170c608, 0x2d5e3354, 0xd4de495a, 0x64c6d006, 0xbcc0c62c, 0x3dd00db3, 0x708f8f34, 0x77d51b42, + 0x264f620f, 0x24b8d2bf, 0x15c1b79e, 0x46a52564, 0xf8d7e54e, 0x3e378160, 0x7895cda5, 0x859c15a5, + 0xe6459788, 0xc37bc75f, 0xdb07ba0c, 0x0676a3ab, 0x7f229b1e, 0x31842e7b, 0x24259fd7, 0xf8bef472, + 0x835ffcb8, 0x6df4c1f2, 0x96f5b195, 0xfd0af0fc, 0xb0fe134c, 0xe2506d3d, 0x4f9b12ea, 0xf215f225, + 0xa223736f, 0x9fb4c428, 0x25d04979, 0x34c713f8, 0xc4618187, 0xea7a6e98, 0x7cd16efc, 0x1436876c, + 0xf1544107, 0xbedeee14, 0x56e9af27, 0xa04aa441, 0x3cf7c899, 0x92ecbae6, 0xdd67016d, 0x151682eb, + 0xa842eedf, 0xfdba60b4, 0xf1907b75, 0x20e3030f, 0x24d8c29e, 0xe139673b, 0xefa63fb8, 0x71873054, + 0xb6f2cf3b, 0x9f326442, 0xcb15a4cc, 0xb01a4504, 0xf1e47d8d, 0x844a1be5, 0xbae7dfdc, 0x42cbda70, + 0xcd7dae0a, 0x57e85b7a, 0xd53f5af6, 0x20cf4d8c, 0xcea4d428, 0x79d130a4, 0x3486ebfb, 0x33d3cddc, + 0x77853b53, 0x37effcb5, 0xc5068778, 0xe580b3e6, 0x4e68b8f4, 0xc5c8b37e, 0x0d809ea2, 0x398feb7c, + 0x132a4f94, 0x43b7950e, 0x2fee7d1c, 0x223613bd, 0xdd06caa2, 0x37df932b, 0xc4248289, 0xacf3ebc3, + 0x5715f6b7, 0xef3478dd, 0xf267616f, 0xc148cbe4, 0x9052815e, 0x5e410fab, 0xb48a2465, 0x2eda7fa4, + 0xe87b40e4, 0xe98ea084, 0x5889e9e1, 0xefd390fc, 0xdd07d35b, 0xdb485694, 0x38d7e5b2, 0x57720101, + 0x730edebc, 0x5b643113, 0x94917e4f, 0x503c2fba, 0x646f1282, 0x7523d24a, 0xe0779695, 0xf9c17a8f, + 0x7a5b2121, 0xd187b896, 0x29263a4d, 0xba510cdf, 0x81f47c9f, 0xad1163ed, 0xea7b5965, 0x1a00726e, + 0x11403092, 0x00da6d77, 0x4a0cdd61, 0xad1f4603, 0x605bdfb0, 0x9eedc364, 0x22ebe6a8, 0xcee7d28a, + 0xa0e736a0, 0x5564a6b9, 0x10853209, 0xc7eb8f37, 0x2de705ca, 0x8951570f, 0xdf09822b, 0xbd691a6c, + 0xaa12e4f2, 0x87451c0f, 0xe0f6a27a, 0x3ada4819, 0x4cf1764f, 0x0d771c2b, 0x67cdb156, 0x350d8384, + 0x5938fa0f, 0x42399ef3, 0x36997b07, 0x0e84093d, 0x4aa93e61, 0x8360d87b, 0x1fa98b0c, 0x1149382c, + 0xe97625a5, 0x0614d1b7, 0x0e25244b, 0x0c768347, 0x589e8d82, 0x0d2059d1, 0xa466bb1e, 0xf8da0a82, + 0x04f19130, 0xba6e4ec0, 0x99265164, 0x1ee7230d, 0x50b2ad80, 0xeaee6801, 0x8db2a283, 0xea8bf59e +}; + +static void generate_round_keys(int rnds, uint32_t* K, uint32_t* x, uint32_t* z) +{ + COMPUTE_Z; + + K[1] = S5[IA(z[2])] ^ S6[IB(z[2])] ^ S7[ID(z[1])] ^ S8[IC(z[1])] ^ S5[IC(z[0])]; + K[2] = S5[IC(z[2])] ^ S6[ID(z[2])] ^ S7[IB(z[1])] ^ S8[IA(z[1])] ^ S6[IC(z[1])]; + K[3] = S5[IA(z[3])] ^ S6[IB(z[3])] ^ S7[ID(z[0])] ^ S8[IC(z[0])] ^ S7[IB(z[2])]; + K[4] = S5[IC(z[3])] ^ S6[ID(z[3])] ^ S7[IB(z[0])] ^ S8[IA(z[0])] ^ S8[IA(z[3])]; + + COMPUTE_X; + + K[5] = S5[ID(x[0])] ^ S6[IC(x[0])] ^ S7[IA(x[3])] ^ S8[IB(x[3])] ^ S5[IA(x[2])]; + K[6] = S5[IB(x[0])] ^ S6[IA(x[0])] ^ S7[IC(x[3])] ^ S8[ID(x[3])] ^ S6[IB(x[3])]; + K[7] = S5[ID(x[1])] ^ S6[IC(x[1])] ^ S7[IA(x[2])] ^ S8[IB(x[2])] ^ S7[ID(x[0])]; + K[8] = S5[IB(x[1])] ^ S6[IA(x[1])] ^ S7[IC(x[2])] ^ S8[ID(x[2])] ^ S8[ID(x[1])]; + + COMPUTE_Z; + + K[9] = S5[ID(z[0])] ^ S6[IC(z[0])] ^ S7[IA(z[3])] ^ S8[IB(z[3])] ^ S5[IB(z[2])]; + K[10] = S5[IB(z[0])] ^ S6[IA(z[0])] ^ S7[IC(z[3])] ^ S8[ID(z[3])] ^ S6[IA(z[3])]; + K[11] = S5[ID(z[1])] ^ S6[IC(z[1])] ^ S7[IA(z[2])] ^ S8[IB(z[2])] ^ S7[IC(z[0])]; + K[12] = S5[IB(z[1])] ^ S6[IA(z[1])] ^ S7[IC(z[2])] ^ S8[ID(z[2])] ^ S8[IC(z[1])]; + + COMPUTE_X; + + if (rnds == 16) { + K[13] = S5[IA(x[2])] ^ S6[IB(x[2])] ^ S7[ID(x[1])] ^ S8[IC(x[1])] ^ S5[ID(x[0])]; + K[14] = S5[IC(x[2])] ^ S6[ID(x[2])] ^ S7[IB(x[1])] ^ S8[IA(x[1])] ^ S6[ID(x[1])]; + K[15] = S5[IA(x[3])] ^ S6[IB(x[3])] ^ S7[ID(x[0])] ^ S8[IC(x[0])] ^ S7[IA(x[2])]; + K[16] = S5[IC(x[3])] ^ S6[ID(x[3])] ^ S7[IB(x[0])] ^ S8[IA(x[0])] ^ S8[IB(x[3])]; + } +} + +static void encipher(AVCAST5* cs, uint8_t* dst, const uint8_t* src) +{ + uint32_t r, l, f, I; + l = AV_RB32(src); + r = AV_RB32(src + 4); + F1(l, r, 1); + F2(r, l, 2); + F3(l, r, 3); + F1(r, l, 4); + F2(l, r, 5); + F3(r, l, 6); + F1(l, r, 7); + F2(r, l, 8); + F3(l, r, 9); + F1(r, l, 10); + F2(l, r, 11); + F3(r, l, 12); + if (cs->rounds == 16) { + F1(l, r, 13); + F2(r, l, 14); + F3(l, r, 15); + F1(r, l, 16); + } + AV_WB32(dst, r); + AV_WB32(dst + 4, l); +} + +static void decipher(AVCAST5* cs, uint8_t* dst, const uint8_t* src, uint8_t *iv) +{ + uint32_t f, I, r, l; + l = AV_RB32(src); + r = AV_RB32(src + 4); + if (cs->rounds == 16) { + F1(l, r, 16); + F3(r, l, 15); + F2(l, r, 14); + F1(r, l, 13); + } + F3(l, r, 12); + F2(r, l, 11); + F1(l, r, 10); + F3(r, l, 9); + F2(l, r, 8); + F1(r, l, 7); + F3(l, r, 6); + F2(r, l, 5); + F1(l, r, 4); + F3(r, l, 3); + F2(l, r, 2); + F1(r, l, 1); + if (iv) { + r ^= AV_RB32(iv); + l ^= AV_RB32(iv + 4); + memcpy(iv, src, 8); + } + AV_WB32(dst, r); + AV_WB32(dst + 4, l); +} + +struct AVCAST5 *av_cast5_alloc(void) +{ + return av_mallocz(sizeof(struct AVCAST5)); +} + +av_cold int av_cast5_init(AVCAST5* cs, const uint8_t *key, int key_bits) +{ + uint8_t newKey[16]; + int i; + uint32_t p[4], q[4]; + if (key_bits % 8 || key_bits < 40 || key_bits > 128) + return AVERROR(EINVAL); + memset(newKey, 0, sizeof(newKey)); + memcpy(newKey, key, key_bits >> 3); + + cs->rounds = key_bits <= 80 ? 12 : 16; + for (i = 0; i < 4; i++) + q[i] = AV_RB32(newKey + (4 * i)); + generate_round_keys(cs->rounds, cs->Km, q, p); + generate_round_keys(cs->rounds, cs->Kr, q, p); + for (i = 0; i <= cs->rounds; i++) + cs->Kr[i] = cs->Kr[i] & 0x1f; + return 0; +} + +void av_cast5_crypt2(AVCAST5* cs, uint8_t* dst, const uint8_t* src, int count, uint8_t *iv, int decrypt) +{ + int i; + while (count--) { + if (decrypt) { + decipher(cs, dst, src, iv); + } else { + if (iv) { + for (i = 0; i < 8; i++) + dst[i] = src[i] ^ iv[i]; + encipher(cs, dst, dst); + memcpy(iv, dst, 8); + } else { + encipher(cs, dst, src); + } + } + src = src + 8; + dst = dst + 8; + } +} +void av_cast5_crypt(AVCAST5* cs, uint8_t* dst, const uint8_t* src, int count, int decrypt) +{ + while (count--) { + if (decrypt){ + decipher(cs, dst, src, NULL); + } else { + encipher(cs, dst, src); + } + src = src + 8; + dst = dst + 8; + } +} diff --git a/libavutil/cast5.h b/libavutil/cast5.h new file mode 100644 index 0000000000..ad5b347e68 --- /dev/null +++ b/libavutil/cast5.h @@ -0,0 +1,80 @@ +/* + * An implementation of the CAST128 algorithm as mentioned in RFC2144 + * Copyright (c) 2014 Supraja Meedinti + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#ifndef AVUTIL_CAST5_H +#define AVUTIL_CAST5_H + +#include <stdint.h> + + +/** + * @file + * @brief Public header for libavutil CAST5 algorithm + * @defgroup lavu_cast5 CAST5 + * @ingroup lavu_crypto + * @{ + */ + +extern const int av_cast5_size; + +struct AVCAST5; + +/** + * Allocate an AVCAST5 context + * To free the struct: av_free(ptr) + */ +struct AVCAST5 *av_cast5_alloc(void); +/** + * Initialize an AVCAST5 context. + * + * @param ctx an AVCAST5 context + * @param key a key of 5,6,...16 bytes used for encryption/decryption + * @param key_bits number of keybits: possible are 40,48,...,128 + * @return 0 on success, less than 0 on failure + */ +int av_cast5_init(struct AVCAST5 *ctx, const uint8_t *key, int key_bits); + +/** + * Encrypt or decrypt a buffer using a previously initialized context, ECB mode only + * + * @param ctx an AVCAST5 context + * @param dst destination array, can be equal to src + * @param src source array, can be equal to dst + * @param count number of 8 byte blocks + * @param decrypt 0 for encryption, 1 for decryption + */ +void av_cast5_crypt(struct AVCAST5 *ctx, uint8_t *dst, const uint8_t *src, int count, int decrypt); + +/** + * Encrypt or decrypt a buffer using a previously initialized context + * + * @param ctx an AVCAST5 context + * @param dst destination array, can be equal to src + * @param src source array, can be equal to dst + * @param count number of 8 byte blocks + * @param iv initialization vector for CBC mode, NULL for ECB mode + * @param decrypt 0 for encryption, 1 for decryption + */ +void av_cast5_crypt2(struct AVCAST5 *ctx, uint8_t *dst, const uint8_t *src, int count, uint8_t *iv, int decrypt); +/** + * @} + */ +#endif /* AVUTIL_CAST5_H */ diff --git a/libavutil/channel_layout.c b/libavutil/channel_layout.c index 41340ecdb6..26c87c96a8 100644 --- a/libavutil/channel_layout.c +++ b/libavutil/channel_layout.c @@ -1,20 +1,20 @@ /* * Copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -28,41 +28,47 @@ #include "avstring.h" #include "avutil.h" #include "channel_layout.h" +#include "bprint.h" #include "common.h" -static const char * const channel_names[] = { - [0] = "FL", /* front left */ - [1] = "FR", /* front right */ - [2] = "FC", /* front center */ - [3] = "LFE", /* low frequency */ - [4] = "BL", /* back left */ - [5] = "BR", /* back right */ - [6] = "FLC", /* front left-of-center */ - [7] = "FRC", /* front right-of-center */ - [8] = "BC", /* back-center */ - [9] = "SL", /* side left */ - [10] = "SR", /* side right */ - [11] = "TC", /* top center */ - [12] = "TFL", /* top front left */ - [13] = "TFC", /* top front center */ - [14] = "TFR", /* top front right */ - [15] = "TBL", /* top back left */ - [16] = "TBC", /* top back center */ - [17] = "TBR", /* top back right */ - [29] = "DL", /* downmix left */ - [30] = "DR", /* downmix right */ - [31] = "WL", /* wide left */ - [32] = "WR", /* wide right */ - [33] = "SDL", /* surround direct left */ - [34] = "SDR", /* surround direct right */ - [35] = "LFE2", /* low frequency 2 */ +struct channel_name { + const char *name; + const char *description; +}; + +static const struct channel_name channel_names[] = { + [0] = { "FL", "front left" }, + [1] = { "FR", "front right" }, + [2] = { "FC", "front center" }, + [3] = { "LFE", "low frequency" }, + [4] = { "BL", "back left" }, + [5] = { "BR", "back right" }, + [6] = { "FLC", "front left-of-center" }, + [7] = { "FRC", "front right-of-center" }, + [8] = { "BC", "back center" }, + [9] = { "SL", "side left" }, + [10] = { "SR", "side right" }, + [11] = { "TC", "top center" }, + [12] = { "TFL", "top front left" }, + [13] = { "TFC", "top front center" }, + [14] = { "TFR", "top front right" }, + [15] = { "TBL", "top back left" }, + [16] = { "TBC", "top back center" }, + [17] = { "TBR", "top back right" }, + [29] = { "DL", "downmix left" }, + [30] = { "DR", "downmix right" }, + [31] = { "WL", "wide left" }, + [32] = { "WR", "wide right" }, + [33] = { "SDL", "surround direct left" }, + [34] = { "SDR", "surround direct right" }, + [35] = { "LFE2", "low frequency 2" }, }; static const char *get_channel_name(int channel_id) { if (channel_id < 0 || channel_id >= FF_ARRAY_ELEMS(channel_names)) return NULL; - return channel_names[channel_id]; + return channel_names[channel_id].name; } static const struct { @@ -72,34 +78,32 @@ static const struct { } channel_layout_map[] = { { "mono", 1, AV_CH_LAYOUT_MONO }, { "stereo", 2, AV_CH_LAYOUT_STEREO }, - { "stereo", 2, AV_CH_LAYOUT_STEREO_DOWNMIX }, { "2.1", 3, AV_CH_LAYOUT_2POINT1 }, { "3.0", 3, AV_CH_LAYOUT_SURROUND }, { "3.0(back)", 3, AV_CH_LAYOUT_2_1 }, - { "3.1", 4, AV_CH_LAYOUT_3POINT1 }, { "4.0", 4, AV_CH_LAYOUT_4POINT0 }, { "quad", 4, AV_CH_LAYOUT_QUAD }, { "quad(side)", 4, AV_CH_LAYOUT_2_2 }, - { "4.1", 5, AV_CH_LAYOUT_4POINT1 }, - { "5.0", 5, AV_CH_LAYOUT_5POINT0 }, + { "3.1", 4, AV_CH_LAYOUT_3POINT1 }, { "5.0", 5, AV_CH_LAYOUT_5POINT0_BACK }, - { "5.1", 6, AV_CH_LAYOUT_5POINT1 }, + { "5.0(side)", 5, AV_CH_LAYOUT_5POINT0 }, + { "4.1", 5, AV_CH_LAYOUT_4POINT1 }, { "5.1", 6, AV_CH_LAYOUT_5POINT1_BACK }, + { "5.1(side)", 6, AV_CH_LAYOUT_5POINT1 }, { "6.0", 6, AV_CH_LAYOUT_6POINT0 }, { "6.0(front)", 6, AV_CH_LAYOUT_6POINT0_FRONT }, { "hexagonal", 6, AV_CH_LAYOUT_HEXAGONAL }, { "6.1", 7, AV_CH_LAYOUT_6POINT1 }, - { "6.1", 7, AV_CH_LAYOUT_6POINT1_BACK }, + { "6.1(back)", 7, AV_CH_LAYOUT_6POINT1_BACK }, { "6.1(front)", 7, AV_CH_LAYOUT_6POINT1_FRONT }, { "7.0", 7, AV_CH_LAYOUT_7POINT0 }, { "7.0(front)", 7, AV_CH_LAYOUT_7POINT0_FRONT }, { "7.1", 8, AV_CH_LAYOUT_7POINT1 }, - { "7.1(wide)", 8, AV_CH_LAYOUT_7POINT1_WIDE }, { "7.1(wide)", 8, AV_CH_LAYOUT_7POINT1_WIDE_BACK }, + { "7.1(wide-side)", 8, AV_CH_LAYOUT_7POINT1_WIDE }, { "octagonal", 8, AV_CH_LAYOUT_OCTAGONAL }, { "hexadecagonal", 16, AV_CH_LAYOUT_HEXADECAGONAL }, { "downmix", 2, AV_CH_LAYOUT_STEREO_DOWNMIX, }, - { 0 } }; static uint64_t get_channel_layout_single(const char *name, int name_len) @@ -108,22 +112,26 @@ static uint64_t get_channel_layout_single(const char *name, int name_len) char *end; int64_t layout; - for (i = 0; i < FF_ARRAY_ELEMS(channel_layout_map) - 1; i++) { + for (i = 0; i < FF_ARRAY_ELEMS(channel_layout_map); i++) { if (strlen(channel_layout_map[i].name) == name_len && !memcmp(channel_layout_map[i].name, name, name_len)) return channel_layout_map[i].layout; } for (i = 0; i < FF_ARRAY_ELEMS(channel_names); i++) - if (channel_names[i] && - strlen(channel_names[i]) == name_len && - !memcmp(channel_names[i], name, name_len)) + if (channel_names[i].name && + strlen(channel_names[i].name) == name_len && + !memcmp(channel_names[i].name, name, name_len)) return (int64_t)1 << i; + + errno = 0; i = strtol(name, &end, 10); - if (end - name == name_len || - (end + 1 - name == name_len && *end == 'c')) + + if (!errno && (end + 1 - name == name_len && *end == 'c')) return av_get_default_channel_layout(i); + + errno = 0; layout = strtoll(name, &end, 0); - if (end - name == name_len) + if (!errno && end - name == name_len) return FFMAX(layout, 0); return 0; } @@ -144,59 +152,60 @@ uint64_t av_get_channel_layout(const char *name) return layout; } -void av_get_channel_layout_string(char *buf, int buf_size, - int nb_channels, uint64_t channel_layout) +void av_bprint_channel_layout(struct AVBPrint *bp, + int nb_channels, uint64_t channel_layout) { int i; if (nb_channels <= 0) nb_channels = av_get_channel_layout_nb_channels(channel_layout); - for (i = 0; channel_layout_map[i].name; i++) + for (i = 0; i < FF_ARRAY_ELEMS(channel_layout_map); i++) if (nb_channels == channel_layout_map[i].nb_channels && channel_layout == channel_layout_map[i].layout) { - av_strlcpy(buf, channel_layout_map[i].name, buf_size); + av_bprintf(bp, "%s", channel_layout_map[i].name); return; } - snprintf(buf, buf_size, "%d channels", nb_channels); + av_bprintf(bp, "%d channels", nb_channels); if (channel_layout) { int i, ch; - av_strlcat(buf, " (", buf_size); + av_bprintf(bp, " ("); for (i = 0, ch = 0; i < 64; i++) { if ((channel_layout & (UINT64_C(1) << i))) { const char *name = get_channel_name(i); if (name) { if (ch > 0) - av_strlcat(buf, "|", buf_size); - av_strlcat(buf, name, buf_size); + av_bprintf(bp, "+"); + av_bprintf(bp, "%s", name); } ch++; } } - av_strlcat(buf, ")", buf_size); + av_bprintf(bp, ")"); } } +void av_get_channel_layout_string(char *buf, int buf_size, + int nb_channels, uint64_t channel_layout) +{ + AVBPrint bp; + + av_bprint_init_for_buffer(&bp, buf, buf_size); + av_bprint_channel_layout(&bp, nb_channels, channel_layout); +} + int av_get_channel_layout_nb_channels(uint64_t channel_layout) { return av_popcount64(channel_layout); } -uint64_t av_get_default_channel_layout(int nb_channels) -{ - switch(nb_channels) { - case 1: return AV_CH_LAYOUT_MONO; - case 2: return AV_CH_LAYOUT_STEREO; - case 3: return AV_CH_LAYOUT_SURROUND; - case 4: return AV_CH_LAYOUT_QUAD; - case 5: return AV_CH_LAYOUT_5POINT0; - case 6: return AV_CH_LAYOUT_5POINT1; - case 7: return AV_CH_LAYOUT_6POINT1; - case 8: return AV_CH_LAYOUT_7POINT1; - case 16: return AV_CH_LAYOUT_HEXADECAGONAL; - default: return 0; - } +int64_t av_get_default_channel_layout(int nb_channels) { + int i; + for (i = 0; i < FF_ARRAY_ELEMS(channel_layout_map); i++) + if (nb_channels == channel_layout_map[i].nb_channels) + return channel_layout_map[i].layout; + return 0; } int av_get_channel_layout_channel_index(uint64_t channel_layout, @@ -220,6 +229,17 @@ const char *av_get_channel_name(uint64_t channel) return NULL; } +const char *av_get_channel_description(uint64_t channel) +{ + int i; + if (av_get_channel_layout_nb_channels(channel) != 1) + return NULL; + for (i = 0; i < FF_ARRAY_ELEMS(channel_names); i++) + if ((1ULL<<i) & channel) + return channel_names[i].description; + return NULL; +} + uint64_t av_channel_layout_extract_channel(uint64_t channel_layout, int index) { int i; @@ -233,3 +253,13 @@ uint64_t av_channel_layout_extract_channel(uint64_t channel_layout, int index) } return 0; } + +int av_get_standard_channel_layout(unsigned index, uint64_t *layout, + const char **name) +{ + if (index >= FF_ARRAY_ELEMS(channel_layout_map)) + return AVERROR_EOF; + if (layout) *layout = channel_layout_map[index].layout; + if (name) *name = channel_layout_map[index].name; + return 0; +} diff --git a/libavutil/channel_layout.h b/libavutil/channel_layout.h index 5bd0c2ce6e..ec7effead1 100644 --- a/libavutil/channel_layout.h +++ b/libavutil/channel_layout.h @@ -2,20 +2,20 @@ * Copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at> * Copyright (c) 2008 Peter Ross * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -36,6 +36,14 @@ /** * @defgroup channel_masks Audio channel masks + * + * A channel layout is a 64-bits integer with a bit set for every channel. + * The number of bits set must be equal to the number of channels. + * The value 0 means that the channel layout is not known. + * @note this data structure is not powerful enough to handle channels + * combinations that have the same channel multiple times, such as + * dual-mono. + * * @{ */ #define AV_CH_FRONT_LEFT 0x00000001 @@ -71,7 +79,7 @@ /** * @} - * @defgroup channel_mask_c Audio channel convenience macros + * @defgroup channel_mask_c Audio channel layouts * @{ * */ #define AV_CH_LAYOUT_MONO (AV_CH_FRONT_CENTER) @@ -115,10 +123,6 @@ enum AVMatrixEncoding { }; /** - * @} - */ - -/** * Return a channel layout id that matches name, or 0 if no match is found. * * name can be one or several of the following notations, @@ -133,7 +137,12 @@ enum AVMatrixEncoding { * - a channel layout mask, in hexadecimal starting with "0x" (see the * AV_CH_* macros). * - * Example: "stereo+FC" = "2+FC" = "2c+1c" = "0x7" + * @warning Starting from the next major bump the trailing character + * 'c' to specify a number of channels will be required, while a + * channel layout mask could also be specified as a decimal number + * (if and only if not followed by "c"). + * + * Example: "stereo+FC" = "2c+FC" = "2c+1c" = "0x7" */ uint64_t av_get_channel_layout(const char *name); @@ -146,6 +155,12 @@ uint64_t av_get_channel_layout(const char *name); */ void av_get_channel_layout_string(char *buf, int buf_size, int nb_channels, uint64_t channel_layout); +struct AVBPrint; +/** + * Append a description of a channel layout to a bprint buffer. + */ +void av_bprint_channel_layout(struct AVBPrint *bp, int nb_channels, uint64_t channel_layout); + /** * Return the number of channels in the channel layout. */ @@ -154,7 +169,7 @@ int av_get_channel_layout_nb_channels(uint64_t channel_layout); /** * Return default channel layout for a given number of channels. */ -uint64_t av_get_default_channel_layout(int nb_channels); +int64_t av_get_default_channel_layout(int nb_channels); /** * Get the index of a channel in channel_layout. @@ -181,6 +196,27 @@ uint64_t av_channel_layout_extract_channel(uint64_t channel_layout, int index); const char *av_get_channel_name(uint64_t channel); /** + * Get the description of a given channel. + * + * @param channel a channel layout with a single channel + * @return channel description on success, NULL on error + */ +const char *av_get_channel_description(uint64_t channel); + +/** + * Get the value and name of a standard channel layout. + * + * @param[in] index index in an internal list, starting at 0 + * @param[out] layout channel layout mask + * @param[out] name name of the layout + * @return 0 if the layout exists, + * <0 if index is beyond the limits + */ +int av_get_standard_channel_layout(unsigned index, uint64_t *layout, + const char **name); + +/** + * @} * @} */ diff --git a/libavutil/color_utils-test.c b/libavutil/color_utils-test.c new file mode 100644 index 0000000000..083872c3e4 --- /dev/null +++ b/libavutil/color_utils-test.c @@ -0,0 +1,43 @@ +/* + * Copyright (c) 2015 Kevin Wheatley <kevin.j.wheatley@gmail.com> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "color_utils.c" + +int main(int argc, char *argv[]) +{ + int i, j; + static const double test_data[] = { + -0.1, -0.018053968510807, -0.01, -0.00449, 0.0, 0.00316227760, 0.005, + 0.009, 0.015, 0.1, 1.0, 52.37, 125.098765, 1999.11123, 6945.443, + 15123.4567, 19845.88923, 98678.4231, 99999.899998 + }; + + for(i = 0; i < AVCOL_TRC_NB; i++) { + avpriv_trc_function func = avpriv_get_trc_function_from_trc(i); + for(j = 0; j < FF_ARRAY_ELEMS(test_data); j++) { + if(func != NULL) { + double result = func(test_data[j]); + printf("AVColorTransferCharacteristic=%d calling func(%f) expected=%f\n", + i, test_data[j], result); + } + } + } + +} diff --git a/libavutil/color_utils.c b/libavutil/color_utils.c new file mode 100644 index 0000000000..63f2230980 --- /dev/null +++ b/libavutil/color_utils.c @@ -0,0 +1,219 @@ +/* + * Copyright (c) 2015 Kevin Wheatley <kevin.j.wheatley@gmail.com> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include <stddef.h> +#include <math.h> + +#include "common.h" +#include "libavutil/color_utils.h" +#include "libavutil/pixfmt.h" + +double avpriv_get_gamma_from_trc(enum AVColorTransferCharacteristic trc) +{ + double gamma; + switch (trc) { + case AVCOL_TRC_BT709: + case AVCOL_TRC_SMPTE170M: + case AVCOL_TRC_SMPTE240M: + case AVCOL_TRC_BT1361_ECG: + case AVCOL_TRC_BT2020_10: + case AVCOL_TRC_BT2020_12: + /* these share a segmented TRC, but gamma 1.961 is a close + approximation, and also more correct for decoding content */ + gamma = 1.961; + break; + case AVCOL_TRC_GAMMA22: + case AVCOL_TRC_IEC61966_2_1: + gamma = 2.2; + break; + case AVCOL_TRC_GAMMA28: + gamma = 2.8; + break; + case AVCOL_TRC_LINEAR: + gamma = 1.0; + break; + default: + gamma = 0.0; // Unknown value representation + } + return gamma; +} + +#define BT709_alpha 1.099296826809442 +#define BT709_beta 0.018053968510807 + +static double avpriv_trc_bt709(double Lc) +{ + const double a = BT709_alpha; + const double b = BT709_beta; + + return (0.0 > Lc) ? 0.0 + : ( b > Lc) ? 4.500 * Lc + : a * pow(Lc, 0.45) - (a - 1.0); +} + +static double avpriv_trc_gamma22(double Lc) +{ + return (0.0 > Lc) ? 0.0 : pow(Lc, 1.0/ 2.2); +} + +static double avpriv_trc_gamma28(double Lc) +{ + return (0.0 > Lc) ? 0.0 : pow(Lc, 1.0/ 2.8); +} + +static double avpriv_trc_smpte240M(double Lc) +{ + const double a = 1.1115; + const double b = 0.0228; + + return (0.0 > Lc) ? 0.0 + : ( b > Lc) ? 4.000 * Lc + : a * pow(Lc, 0.45) - (a - 1.0); +} + +static double avpriv_trc_linear(double Lc) +{ + return Lc; +} + +static double avpriv_trc_log(double Lc) +{ + return (0.01 > Lc) ? 0.0 : 1.0 + log10(Lc) / 2.0; +} + +static double avpriv_trc_log_sqrt(double Lc) +{ + // sqrt(10) / 1000 + return (0.00316227766 > Lc) ? 0.0 : 1.0 + log10(Lc) / 2.5; +} + +static double avpriv_trc_iec61966_2_4(double Lc) +{ + const double a = BT709_alpha; + const double b = BT709_beta; + + return (-b >= Lc) ? -a * pow(-Lc, 0.45) + (a - 1.0) + : ( b > Lc) ? 4.500 * Lc + : a * pow( Lc, 0.45) - (a - 1.0); +} + +static double avpriv_trc_bt1361(double Lc) +{ + const double a = BT709_alpha; + const double b = BT709_beta; + + return (-0.0045 >= Lc) ? -(a * pow(-4.0 * Lc, 0.45) + (a - 1.0)) / 4.0 + : ( b > Lc) ? 4.500 * Lc + : a * pow( Lc, 0.45) - (a - 1.0); +} + +static double avpriv_trc_iec61966_2_1(double Lc) +{ + const double a = 1.055; + const double b = 0.0031308; + + return (0.0 > Lc) ? 0.0 + : ( b > Lc) ? 12.92 * Lc + : a * pow(Lc, 1.0 / 2.4) - (a - 1.0); +} + +static double avpriv_trc_smpte_st2084(double Lc) +{ + const double c1 = 3424.0 / 4096.0; // c3-c2 + 1 + const double c2 = 32.0 * 2413.0 / 4096.0; + const double c3 = 32.0 * 2392.0 / 4096.0; + const double m = 128.0 * 2523.0 / 4096.0; + const double n = 0.25 * 2610.0 / 4096.0; + const double L = Lc / 10000.0; + const double Ln = pow(L, n); + + return (0.0 > Lc) ? 0.0 + : pow((c1 + c2 * Ln) / (1.0 + c3 * Ln), m); + +} + +static double avpriv_trc_smpte_st428_1(double Lc) +{ + return (0.0 > Lc) ? 0.0 + : pow(48.0 * Lc / 52.37, 1.0 / 2.6); +} + +avpriv_trc_function avpriv_get_trc_function_from_trc(enum AVColorTransferCharacteristic trc) +{ + avpriv_trc_function func = NULL; + switch (trc) { + case AVCOL_TRC_BT709: + case AVCOL_TRC_SMPTE170M: + case AVCOL_TRC_BT2020_10: + case AVCOL_TRC_BT2020_12: + func = avpriv_trc_bt709; + break; + + case AVCOL_TRC_GAMMA22: + func = avpriv_trc_gamma22; + break; + case AVCOL_TRC_GAMMA28: + func = avpriv_trc_gamma28; + break; + + case AVCOL_TRC_SMPTE240M: + func = avpriv_trc_smpte240M; + break; + + case AVCOL_TRC_LINEAR: + func = avpriv_trc_linear; + break; + + case AVCOL_TRC_LOG: + func = avpriv_trc_log; + break; + + case AVCOL_TRC_LOG_SQRT: + func = avpriv_trc_log_sqrt; + break; + + case AVCOL_TRC_IEC61966_2_4: + func = avpriv_trc_iec61966_2_4; + break; + + case AVCOL_TRC_BT1361_ECG: + func = avpriv_trc_bt1361; + break; + + case AVCOL_TRC_IEC61966_2_1: + func = avpriv_trc_iec61966_2_1; + break; + + case AVCOL_TRC_SMPTEST2084: + func = avpriv_trc_smpte_st2084; + break; + + case AVCOL_TRC_SMPTEST428_1: + func = avpriv_trc_smpte_st428_1; + break; + + case AVCOL_TRC_RESERVED0: + case AVCOL_TRC_UNSPECIFIED: + case AVCOL_TRC_RESERVED: + default: + break; + } + return func; +} diff --git a/libavutil/color_utils.h b/libavutil/color_utils.h new file mode 100644 index 0000000000..9529006452 --- /dev/null +++ b/libavutil/color_utils.h @@ -0,0 +1,56 @@ +/* + * Copyright (c) 2015 Kevin Wheatley <kevin.j.wheatley@gmail.com> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#ifndef AVUTIL_COLOR_UTILS_H +#define AVUTIL_COLOR_UTILS_H + + +#include "libavutil/pixfmt.h" + +/** + * Determine a suitable 'gamma' value to match the supplied + * AVColorTransferCharacteristic. + * + * See Apple Technical Note TN2257 (https://developer.apple.com/library/mac/technotes/tn2257/_index.html) + * + * @return Will return an approximation to the simple gamma function matching + * the supplied Transfer Characteristic, Will return 0.0 for any + * we cannot reasonably match against. + */ +double avpriv_get_gamma_from_trc(enum AVColorTransferCharacteristic trc); + + +typedef double (*avpriv_trc_function)(double); + +/** + * Determine the function needed to apply the given + * AVColorTransferCharacteristic to linear input. + * + * The function returned should expect a nominal domain and range of [0.0-1.0] + * values outside of this range maybe valid depending on the chosen + * characteristic function. + * + * @return Will return pointer to the function matching the + * supplied Transfer Characteristic. If unspecified will + * return NULL: + */ +avpriv_trc_function avpriv_get_trc_function_from_trc(enum AVColorTransferCharacteristic trc); + +#endif diff --git a/libavutil/colorspace.h b/libavutil/colorspace.h index 8757566c78..7d3f7110c9 100644 --- a/libavutil/colorspace.h +++ b/libavutil/colorspace.h @@ -2,20 +2,20 @@ * Colorspace conversion defines * Copyright (c) 2001, 2002, 2003 Fabrice Bellard * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -41,6 +41,16 @@ b_add = FIX(1.77200*255.0/224.0) * cb + ONE_HALF;\ } +#define YUV_TO_RGB1_CCIR_BT709(cb1, cr1)\ +{\ + cb = (cb1) - 128;\ + cr = (cr1) - 128;\ + r_add = FIX(1.5747*255.0/224.0) * cr + ONE_HALF;\ + g_add = - FIX(0.1873*255.0/224.0) * cb - FIX(0.4682*255.0/224.0) * cr + \ + ONE_HALF;\ + b_add = FIX(1.8556*255.0/224.0) * cb + ONE_HALF;\ +} + #define YUV_TO_RGB2_CCIR(r, g, b, y1)\ {\ y = ((y1) - 16) * FIX(255.0/219.0);\ @@ -84,18 +94,6 @@ static inline int C_JPEG_TO_CCIR(int y) { } -#define RGB_TO_Y(r, g, b) \ -((FIX(0.29900) * (r) + FIX(0.58700) * (g) + \ - FIX(0.11400) * (b) + ONE_HALF) >> SCALEBITS) - -#define RGB_TO_U(r1, g1, b1, shift)\ -(((- FIX(0.16874) * r1 - FIX(0.33126) * g1 + \ - FIX(0.50000) * b1 + (ONE_HALF << shift) - 1) >> (SCALEBITS + shift)) + 128) - -#define RGB_TO_V(r1, g1, b1, shift)\ -(((FIX(0.50000) * r1 - FIX(0.41869) * g1 - \ - FIX(0.08131) * b1 + (ONE_HALF << shift) - 1) >> (SCALEBITS + shift)) + 128) - #define RGB_TO_Y_CCIR(r, g, b) \ ((FIX(0.29900*219.0/255.0) * (r) + FIX(0.58700*219.0/255.0) * (g) + \ FIX(0.11400*219.0/255.0) * (b) + (ONE_HALF + (16 << SCALEBITS))) >> SCALEBITS) diff --git a/libavutil/common.h b/libavutil/common.h index d2ddcba9ed..8142b31fdb 100644 --- a/libavutil/common.h +++ b/libavutil/common.h @@ -1,20 +1,20 @@ /* * copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -26,6 +26,10 @@ #ifndef AVUTIL_COMMON_H #define AVUTIL_COMMON_H +#if defined(__cplusplus) && !defined(__STDC_CONSTANT_MACROS) && !defined(UINT64_C) +#error missing -D__STDC_CONSTANT_MACROS / #define __STDC_CONSTANT_MACROS +#endif + #include <errno.h> #include <inttypes.h> #include <limits.h> @@ -50,17 +54,43 @@ #define RSHIFT(a,b) ((a) > 0 ? ((a) + ((1<<(b))>>1))>>(b) : ((a) + ((1<<(b))>>1)-1)>>(b)) /* assume b>0 */ #define ROUNDED_DIV(a,b) (((a)>0 ? (a) + ((b)>>1) : (a) - ((b)>>1))/(b)) +/* Fast a/(1<<b) rounded toward +inf. Assume a>=0 and b>=0 */ +#define AV_CEIL_RSHIFT(a,b) (!av_builtin_constant_p(b) ? -((-(a)) >> (b)) \ + : ((a) + (1<<(b)) - 1) >> (b)) +/* Backwards compat. */ +#define FF_CEIL_RSHIFT AV_CEIL_RSHIFT + +#define FFUDIV(a,b) (((a)>0 ?(a):(a)-(b)+1) / (b)) +#define FFUMOD(a,b) ((a)-(b)*FFUDIV(a,b)) /** - * Fast a / (1 << b) rounded toward +inf, assuming a >= 0 and b >= 0. + * Absolute value, Note, INT_MIN / INT64_MIN result in undefined behavior as they + * are not representable as absolute values of their type. This is the same + * as with *abs() + * @see FFNABS() */ -#define AV_CEIL_RSHIFT(a, b) \ - (av_builtin_constant_p(b) ? ((a) + (1 << (b)) - 1) >> (b) \ - : -((-(a)) >> (b))) - #define FFABS(a) ((a) >= 0 ? (a) : (-(a))) #define FFSIGN(a) ((a) > 0 ? 1 : -1) +/** + * Negative Absolute value. + * this works for all integers of all types. + * As with many macros, this evaluates its argument twice, it thus must not have + * a sideeffect, that is FFNABS(x++) has undefined behavior. + */ +#define FFNABS(a) ((a) <= 0 ? (a) : (-(a))) + +/** + * Comparator. + * For two numerical expressions x and y, gives 1 if x > y, -1 if x < y, and 0 + * if x == y. This is useful for instance in a qsort comparator callback. + * Furthermore, compilers are able to optimize this to branchless code, and + * there is no risk of overflow with signed types. + * As with many macros, this evaluates its argument multiple times, it thus + * must not have a side-effect. + */ +#define FFDIFFSIGN(x,y) (((x)>(y)) - ((x)<(y))) + #define FFMAX(a,b) ((a) > (b) ? (a) : (b)) #define FFMAX3(a,b,c) FFMAX(FFMAX(a,b),c) #define FFMIN(a,b) ((a) > (b) ? (b) : (a)) @@ -96,6 +126,26 @@ av_const int av_log2_16bit(unsigned v); */ static av_always_inline av_const int av_clip_c(int a, int amin, int amax) { +#if defined(HAVE_AV_CONFIG_H) && defined(ASSERT_LEVEL) && ASSERT_LEVEL >= 2 + if (amin > amax) abort(); +#endif + if (a < amin) return amin; + else if (a > amax) return amax; + else return a; +} + +/** + * Clip a signed 64bit integer value into the amin-amax range. + * @param a value to clip + * @param amin minimum value of the clip range + * @param amax maximum value of the clip range + * @return clipped value + */ +static av_always_inline av_const int64_t av_clip64_c(int64_t a, int64_t amin, int64_t amax) +{ +#if defined(HAVE_AV_CONFIG_H) && defined(ASSERT_LEVEL) && ASSERT_LEVEL >= 2 + if (amin > amax) abort(); +#endif if (a < amin) return amin; else if (a > amax) return amax; else return a; @@ -119,7 +169,7 @@ static av_always_inline av_const uint8_t av_clip_uint8_c(int a) */ static av_always_inline av_const int8_t av_clip_int8_c(int a) { - if ((a+0x80) & ~0xFF) return (a>>31) ^ 0x7F; + if ((a+0x80U) & ~0xFF) return (a>>31) ^ 0x7F; else return a; } @@ -141,7 +191,7 @@ static av_always_inline av_const uint16_t av_clip_uint16_c(int a) */ static av_always_inline av_const int16_t av_clip_int16_c(int a) { - if ((a+0x8000) & ~0xFFFF) return (a>>31) ^ 0x7FFF; + if ((a+0x8000U) & ~0xFFFF) return (a>>31) ^ 0x7FFF; else return a; } @@ -152,8 +202,8 @@ static av_always_inline av_const int16_t av_clip_int16_c(int a) */ static av_always_inline av_const int32_t av_clipl_int32_c(int64_t a) { - if ((a+0x80000000u) & ~UINT64_C(0xFFFFFFFF)) return (a>>63) ^ 0x7FFFFFFF; - else return a; + if ((a+0x80000000u) & ~UINT64_C(0xFFFFFFFF)) return (int32_t)((a>>63) ^ 0x7FFFFFFF); + else return (int32_t)a; } /** @@ -164,7 +214,7 @@ static av_always_inline av_const int32_t av_clipl_int32_c(int64_t a) */ static av_always_inline av_const int av_clip_intp2_c(int a, int p) { - if ((a + (1 << p)) & ~((1 << (p + 1)) - 1)) + if (((unsigned)a + (1 << p)) & ~((2 << p) - 1)) return (a >> 31) ^ ((1 << p) - 1); else return a; @@ -183,6 +233,17 @@ static av_always_inline av_const unsigned av_clip_uintp2_c(int a, int p) } /** + * Clear high bits from an unsigned integer starting with specific bit position + * @param a value to clip + * @param p bit position to clip at + * @return clipped value + */ +static av_always_inline av_const unsigned av_mod_uintp2_c(unsigned a, unsigned p) +{ + return a & ((1 << p) - 1); +} + +/** * Add two signed 32-bit values with saturation. * * @param a one value @@ -215,6 +276,26 @@ static av_always_inline int av_sat_dadd32_c(int a, int b) */ static av_always_inline av_const float av_clipf_c(float a, float amin, float amax) { +#if defined(HAVE_AV_CONFIG_H) && defined(ASSERT_LEVEL) && ASSERT_LEVEL >= 2 + if (amin > amax) abort(); +#endif + if (a < amin) return amin; + else if (a > amax) return amax; + else return a; +} + +/** + * Clip a double value into the amin-amax range. + * @param a value to clip + * @param amin minimum value of the clip range + * @param amax maximum value of the clip range + * @return clipped value + */ +static av_always_inline av_const double av_clipd_c(double a, double amin, double amax) +{ +#if defined(HAVE_AV_CONFIG_H) && defined(ASSERT_LEVEL) && ASSERT_LEVEL >= 2 + if (amin > amax) abort(); +#endif if (a < amin) return amin; else if (a > amax) return amax; else return a; @@ -250,7 +331,12 @@ static av_always_inline av_const int av_popcount_c(uint32_t x) */ static av_always_inline av_const int av_popcount64_c(uint64_t x) { - return av_popcount(x) + av_popcount(x >> 32); + return av_popcount((uint32_t)x) + av_popcount((uint32_t)(x >> 32)); +} + +static av_always_inline av_const int av_parity_c(uint32_t v) +{ + return av_popcount(v) & 1; } #define MKTAG(a,b,c,d) ((a) | ((b) << 8) | ((c) << 16) | ((unsigned)(d) << 24)) @@ -266,15 +352,20 @@ static av_always_inline av_const int av_popcount64_c(uint64_t x) * input, this could be *ptr++. * @param ERROR Expression to be evaluated on invalid input, * typically a goto statement. + * + * @warning ERROR should not contain a loop control statement which + * could interact with the internal while loop, and should force an + * exit from the macro code (e.g. through a goto or a return) in order + * to prevent undefined results. */ #define GET_UTF8(val, GET_BYTE, ERROR)\ - val= GET_BYTE;\ + val= (GET_BYTE);\ {\ uint32_t top = (val & 128) >> 1;\ - if ((val & 0xc0) == 0x80)\ + if ((val & 0xc0) == 0x80 || val >= 0xFE)\ ERROR\ while (val & top) {\ - int tmp= GET_BYTE - 128;\ + int tmp= (GET_BYTE) - 128;\ if(tmp>>6)\ ERROR\ val= (val<<6) + tmp;\ @@ -389,6 +480,9 @@ static av_always_inline av_const int av_popcount64_c(uint64_t x) #ifndef av_clip # define av_clip av_clip_c #endif +#ifndef av_clip64 +# define av_clip64 av_clip64_c +#endif #ifndef av_clip_uint8 # define av_clip_uint8 av_clip_uint8_c #endif @@ -410,6 +504,9 @@ static av_always_inline av_const int av_popcount64_c(uint64_t x) #ifndef av_clip_uintp2 # define av_clip_uintp2 av_clip_uintp2_c #endif +#ifndef av_mod_uintp2 +# define av_mod_uintp2 av_mod_uintp2_c +#endif #ifndef av_sat_add32 # define av_sat_add32 av_sat_add32_c #endif @@ -419,9 +516,15 @@ static av_always_inline av_const int av_popcount64_c(uint64_t x) #ifndef av_clipf # define av_clipf av_clipf_c #endif +#ifndef av_clipd +# define av_clipd av_clipd_c +#endif #ifndef av_popcount # define av_popcount av_popcount_c #endif #ifndef av_popcount64 # define av_popcount64 av_popcount64_c #endif +#ifndef av_parity +# define av_parity av_parity_c +#endif diff --git a/libavutil/cpu-test.c b/libavutil/cpu-test.c index a2c3b38c0d..3eca6d2033 100644 --- a/libavutil/cpu-test.c +++ b/libavutil/cpu-test.c @@ -1,26 +1,25 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ -#include "config.h" +#include <stdio.h> +#include "avstring.h" -#if HAVE_UNISTD_H -#include <unistd.h> -#elif !HAVE_GETOPT +#if !HAVE_GETOPT #include "compat/getopt.c" #endif @@ -47,6 +46,7 @@ static const struct { { AV_CPU_FLAG_VFP_VM, "vfp_vm" }, { AV_CPU_FLAG_VFPV3, "vfpv3" }, { AV_CPU_FLAG_NEON, "neon" }, + { AV_CPU_FLAG_SETEND, "setend" }, #elif ARCH_PPC { AV_CPU_FLAG_ALTIVEC, "altivec" }, #elif ARCH_X86 @@ -54,9 +54,9 @@ static const struct { { AV_CPU_FLAG_MMXEXT, "mmxext" }, { AV_CPU_FLAG_SSE, "sse" }, { AV_CPU_FLAG_SSE2, "sse2" }, - { AV_CPU_FLAG_SSE2SLOW, "sse2(slow)" }, + { AV_CPU_FLAG_SSE2SLOW, "sse2slow" }, { AV_CPU_FLAG_SSE3, "sse3" }, - { AV_CPU_FLAG_SSE3SLOW, "sse3(slow)" }, + { AV_CPU_FLAG_SSE3SLOW, "sse3slow" }, { AV_CPU_FLAG_SSSE3, "ssse3" }, { AV_CPU_FLAG_ATOM, "atom" }, { AV_CPU_FLAG_SSE4, "sse4.1" }, @@ -72,6 +72,7 @@ static const struct { { AV_CPU_FLAG_AVX2, "avx2" }, { AV_CPU_FLAG_BMI1, "bmi1" }, { AV_CPU_FLAG_BMI2, "bmi2" }, + { AV_CPU_FLAG_AESNI, "aesni" }, #endif { 0 } }; @@ -80,12 +81,12 @@ static void print_cpu_flags(int cpu_flags, const char *type) { int i; - fprintf(stderr, "cpu_flags(%s) = 0x%08X\n", type, cpu_flags); - fprintf(stderr, "cpu_flags_str(%s) =", type); + printf("cpu_flags(%s) = 0x%08X\n", type, cpu_flags); + printf("cpu_flags_str(%s) =", type); for (i = 0; cpu_flag_tab[i].flag; i++) if (cpu_flags & cpu_flag_tab[i].flag) - fprintf(stderr, " %s", cpu_flag_tab[i].name); - fprintf(stderr, "\n"); + printf(" %s", cpu_flag_tab[i].name); + printf("\n"); } @@ -95,6 +96,15 @@ int main(int argc, char **argv) int cpu_flags_eff; int cpu_count = av_cpu_count(); char threads[5] = "auto"; + int i; + + for(i = 0; cpu_flag_tab[i].flag; i++) { + unsigned tmp = 0; + if (av_parse_cpu_caps(&tmp, cpu_flag_tab[i].name) < 0) { + fprintf(stderr, "Table missing %s\n", cpu_flag_tab[i].name); + return 4; + } + } if (cpu_flags_raw < 0) return 1; @@ -106,10 +116,11 @@ int main(int argc, char **argv) switch (c) { case 'c': { - int cpuflags = av_parse_cpu_flags(optarg); - if (cpuflags < 0) + unsigned flags = av_get_cpu_flags(); + if (av_parse_cpu_caps(&flags, optarg) < 0) return 2; - av_set_cpu_flags_mask(cpuflags); + + av_force_cpu_flags(flags); break; } case 't': @@ -130,7 +141,7 @@ int main(int argc, char **argv) print_cpu_flags(cpu_flags_raw, "raw"); print_cpu_flags(cpu_flags_eff, "effective"); - fprintf(stderr, "threads = %s (cpu_count = %d)\n", threads, cpu_count); + printf("threads = %s (cpu_count = %d)\n", threads, cpu_count); return 0; } diff --git a/libavutil/cpu.c b/libavutil/cpu.c index 7d7390a756..f5785fc13f 100644 --- a/libavutil/cpu.c +++ b/libavutil/cpu.c @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -25,10 +25,12 @@ #include "common.h" #if HAVE_SCHED_GETAFFINITY -#define _GNU_SOURCE +#ifndef _GNU_SOURCE +# define _GNU_SOURCE +#endif #include <sched.h> #endif -#if HAVE_GETPROCESSAFFINITYMASK +#if HAVE_GETPROCESSAFFINITYMASK || HAVE_WINRT #include <windows.h> #endif #if HAVE_SYSCTL @@ -38,16 +40,41 @@ #include <sys/types.h> #include <sys/sysctl.h> #endif -#if HAVE_SYSCONF +#if HAVE_UNISTD_H #include <unistd.h> #endif -static int cpuflags_mask = -1, checked; +static int flags, checked; + +void av_force_cpu_flags(int arg){ + if ( (arg & ( AV_CPU_FLAG_3DNOW | + AV_CPU_FLAG_3DNOWEXT | + AV_CPU_FLAG_MMXEXT | + AV_CPU_FLAG_SSE | + AV_CPU_FLAG_SSE2 | + AV_CPU_FLAG_SSE2SLOW | + AV_CPU_FLAG_SSE3 | + AV_CPU_FLAG_SSE3SLOW | + AV_CPU_FLAG_SSSE3 | + AV_CPU_FLAG_SSE4 | + AV_CPU_FLAG_SSE42 | + AV_CPU_FLAG_AVX | + AV_CPU_FLAG_AVXSLOW | + AV_CPU_FLAG_XOP | + AV_CPU_FLAG_FMA3 | + AV_CPU_FLAG_FMA4 | + AV_CPU_FLAG_AVX2 )) + && !(arg & AV_CPU_FLAG_MMX)) { + av_log(NULL, AV_LOG_WARNING, "MMX implied by specified flags\n"); + arg |= AV_CPU_FLAG_MMX; + } + + flags = arg; + checked = arg != -1; +} int av_get_cpu_flags(void) { - static int flags; - if (checked) return flags; @@ -60,16 +87,15 @@ int av_get_cpu_flags(void) if (ARCH_X86) flags = ff_get_cpu_flags_x86(); - flags &= cpuflags_mask; checked = 1; - return flags; } void av_set_cpu_flags_mask(int mask) { - cpuflags_mask = mask; checked = 0; + flags = av_get_cpu_flags() & mask; + checked = 1; } int av_parse_cpu_flags(const char *s) @@ -92,6 +118,7 @@ int av_parse_cpu_flags(const char *s) #define CPUFLAG_FMA4 (AV_CPU_FLAG_FMA4 | CPUFLAG_AVX) #define CPUFLAG_AVX2 (AV_CPU_FLAG_AVX2 | CPUFLAG_AVX) #define CPUFLAG_BMI2 (AV_CPU_FLAG_BMI2 | AV_CPU_FLAG_BMI1) +#define CPUFLAG_AESNI (AV_CPU_FLAG_AESNI | CPUFLAG_SSE42) static const AVOption cpuflags_opts[] = { { "flags" , NULL, 0, AV_OPT_TYPE_FLAGS, { .i64 = 0 }, INT64_MIN, INT64_MAX, .unit = "flags" }, #if ARCH_PPC @@ -119,6 +146,7 @@ int av_parse_cpu_flags(const char *s) { "3dnow" , NULL, 0, AV_OPT_TYPE_CONST, { .i64 = CPUFLAG_3DNOW }, .unit = "flags" }, { "3dnowext", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = CPUFLAG_3DNOWEXT }, .unit = "flags" }, { "cmov", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_CMOV }, .unit = "flags" }, + { "aesni" , NULL, 0, AV_OPT_TYPE_CONST, { .i64 = CPUFLAG_AESNI }, .unit = "flags" }, #elif ARCH_ARM { "armv5te", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_ARMV5TE }, .unit = "flags" }, { "armv6", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_ARMV6 }, .unit = "flags" }, @@ -150,9 +178,89 @@ int av_parse_cpu_flags(const char *s) return flags & INT_MAX; } +int av_parse_cpu_caps(unsigned *flags, const char *s) +{ + static const AVOption cpuflags_opts[] = { + { "flags" , NULL, 0, AV_OPT_TYPE_FLAGS, { .i64 = 0 }, INT64_MIN, INT64_MAX, .unit = "flags" }, +#if ARCH_PPC + { "altivec" , NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_ALTIVEC }, .unit = "flags" }, +#elif ARCH_X86 + { "mmx" , NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_MMX }, .unit = "flags" }, + { "mmx2" , NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_MMX2 }, .unit = "flags" }, + { "mmxext" , NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_MMX2 }, .unit = "flags" }, + { "sse" , NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_SSE }, .unit = "flags" }, + { "sse2" , NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_SSE2 }, .unit = "flags" }, + { "sse2slow", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_SSE2SLOW }, .unit = "flags" }, + { "sse3" , NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_SSE3 }, .unit = "flags" }, + { "sse3slow", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_SSE3SLOW }, .unit = "flags" }, + { "ssse3" , NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_SSSE3 }, .unit = "flags" }, + { "atom" , NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_ATOM }, .unit = "flags" }, + { "sse4.1" , NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_SSE4 }, .unit = "flags" }, + { "sse4.2" , NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_SSE42 }, .unit = "flags" }, + { "avx" , NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_AVX }, .unit = "flags" }, + { "avxslow" , NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_AVXSLOW }, .unit = "flags" }, + { "xop" , NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_XOP }, .unit = "flags" }, + { "fma3" , NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_FMA3 }, .unit = "flags" }, + { "fma4" , NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_FMA4 }, .unit = "flags" }, + { "avx2" , NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_AVX2 }, .unit = "flags" }, + { "bmi1" , NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_BMI1 }, .unit = "flags" }, + { "bmi2" , NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_BMI2 }, .unit = "flags" }, + { "3dnow" , NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_3DNOW }, .unit = "flags" }, + { "3dnowext", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_3DNOWEXT }, .unit = "flags" }, + { "cmov", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_CMOV }, .unit = "flags" }, + { "aesni", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_AESNI }, .unit = "flags" }, + +#define CPU_FLAG_P2 AV_CPU_FLAG_CMOV | AV_CPU_FLAG_MMX +#define CPU_FLAG_P3 CPU_FLAG_P2 | AV_CPU_FLAG_MMX2 | AV_CPU_FLAG_SSE +#define CPU_FLAG_P4 CPU_FLAG_P3| AV_CPU_FLAG_SSE2 + { "pentium2", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = CPU_FLAG_P2 }, .unit = "flags" }, + { "pentium3", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = CPU_FLAG_P3 }, .unit = "flags" }, + { "pentium4", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = CPU_FLAG_P4 }, .unit = "flags" }, + +#define CPU_FLAG_K62 AV_CPU_FLAG_MMX | AV_CPU_FLAG_3DNOW +#define CPU_FLAG_ATHLON CPU_FLAG_K62 | AV_CPU_FLAG_CMOV | AV_CPU_FLAG_3DNOWEXT | AV_CPU_FLAG_MMX2 +#define CPU_FLAG_ATHLONXP CPU_FLAG_ATHLON | AV_CPU_FLAG_SSE +#define CPU_FLAG_K8 CPU_FLAG_ATHLONXP | AV_CPU_FLAG_SSE2 + { "k6", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_MMX }, .unit = "flags" }, + { "k62", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = CPU_FLAG_K62 }, .unit = "flags" }, + { "athlon", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = CPU_FLAG_ATHLON }, .unit = "flags" }, + { "athlonxp", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = CPU_FLAG_ATHLONXP }, .unit = "flags" }, + { "k8", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = CPU_FLAG_K8 }, .unit = "flags" }, +#elif ARCH_ARM + { "armv5te", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_ARMV5TE }, .unit = "flags" }, + { "armv6", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_ARMV6 }, .unit = "flags" }, + { "armv6t2", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_ARMV6T2 }, .unit = "flags" }, + { "vfp", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_VFP }, .unit = "flags" }, + { "vfp_vm", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_VFP_VM }, .unit = "flags" }, + { "vfpv3", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_VFPV3 }, .unit = "flags" }, + { "neon", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_NEON }, .unit = "flags" }, + { "setend", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_SETEND }, .unit = "flags" }, +#elif ARCH_AARCH64 + { "armv8", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_ARMV8 }, .unit = "flags" }, + { "neon", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_NEON }, .unit = "flags" }, + { "vfp", NULL, 0, AV_OPT_TYPE_CONST, { .i64 = AV_CPU_FLAG_VFP }, .unit = "flags" }, +#endif + { NULL }, + }; + static const AVClass class = { + .class_name = "cpuflags", + .item_name = av_default_item_name, + .option = cpuflags_opts, + .version = LIBAVUTIL_VERSION_INT, + }; + const AVClass *pclass = &class; + + return av_opt_eval_flags(&pclass, &cpuflags_opts[0], s, flags); +} + int av_cpu_count(void) { + static volatile int printed; + int nb_cpus = 1; +#if HAVE_WINRT + SYSTEM_INFO sysinfo; +#endif #if HAVE_SCHED_GETAFFINITY && defined(CPU_COUNT) cpu_set_t cpuset; @@ -174,7 +282,15 @@ int av_cpu_count(void) nb_cpus = sysconf(_SC_NPROC_ONLN); #elif HAVE_SYSCONF && defined(_SC_NPROCESSORS_ONLN) nb_cpus = sysconf(_SC_NPROCESSORS_ONLN); +#elif HAVE_WINRT + GetNativeSystemInfo(&sysinfo); + nb_cpus = sysinfo.dwNumberOfProcessors; #endif + if (!printed) { + av_log(NULL, AV_LOG_DEBUG, "detected %d logical cores\n", nb_cpus); + printed = 1; + } + return nb_cpus; } diff --git a/libavutil/cpu.h b/libavutil/cpu.h index d640e79783..4bff16714a 100644 --- a/libavutil/cpu.h +++ b/libavutil/cpu.h @@ -1,33 +1,34 @@ /* * Copyright (c) 2000, 2001, 2002 Fabrice Bellard * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ #ifndef AVUTIL_CPU_H #define AVUTIL_CPU_H -#include "version.h" +#include "attributes.h" #define AV_CPU_FLAG_FORCE 0x80000000 /* force usage of selected flags (OR) */ /* lower 16 bits - CPU features */ #define AV_CPU_FLAG_MMX 0x0001 ///< standard MMX #define AV_CPU_FLAG_MMXEXT 0x0002 ///< SSE integer functions or AMD MMX ext +#define AV_CPU_FLAG_MMX2 0x0002 ///< SSE integer functions or AMD MMX ext #define AV_CPU_FLAG_3DNOW 0x0004 ///< AMD 3DNOW #define AV_CPU_FLAG_SSE 0x0008 ///< SSE functions #define AV_CPU_FLAG_SSE2 0x0010 ///< PIV SSE2 functions @@ -41,11 +42,12 @@ #define AV_CPU_FLAG_ATOM 0x10000000 ///< Atom processor, some SSSE3 instructions are slower #define AV_CPU_FLAG_SSE4 0x0100 ///< Penryn SSE4.1 functions #define AV_CPU_FLAG_SSE42 0x0200 ///< Nehalem SSE4.2 functions +#define AV_CPU_FLAG_AESNI 0x80000 ///< Advanced Encryption Standard functions #define AV_CPU_FLAG_AVX 0x4000 ///< AVX functions: requires OS support even if YMM registers aren't used #define AV_CPU_FLAG_AVXSLOW 0x8000000 ///< AVX supported, but slow when using YMM registers (e.g. Bulldozer) #define AV_CPU_FLAG_XOP 0x0400 ///< Bulldozer XOP functions #define AV_CPU_FLAG_FMA4 0x0800 ///< Bulldozer FMA4 functions -#define AV_CPU_FLAG_CMOV 0x1000 ///< i686 cmov +#define AV_CPU_FLAG_CMOV 0x1000 ///< supports cmov instruction #define AV_CPU_FLAG_AVX2 0x8000 ///< AVX2 functions: requires OS support even if YMM registers aren't used #define AV_CPU_FLAG_FMA3 0x10000 ///< Haswell FMA3 functions #define AV_CPU_FLAG_BMI1 0x20000 ///< Bit Manipulation Instruction Set 1 @@ -63,28 +65,51 @@ #define AV_CPU_FLAG_NEON (1 << 5) #define AV_CPU_FLAG_ARMV8 (1 << 6) #define AV_CPU_FLAG_VFP_VM (1 << 7) ///< VFPv2 vector mode, deprecated in ARMv7-A and unavailable in various CPUs implementations +#define AV_CPU_FLAG_SETEND (1 <<16) /** * Return the flags which specify extensions supported by the CPU. + * The returned value is affected by av_force_cpu_flags() if that was used + * before. So av_get_cpu_flags() can easily be used in an application to + * detect the enabled cpu flags. */ int av_get_cpu_flags(void); /** + * Disables cpu detection and forces the specified flags. + * -1 is a special case that disables forcing of specific flags. + */ +void av_force_cpu_flags(int flags); + +/** * Set a mask on flags returned by av_get_cpu_flags(). * This function is mainly useful for testing. + * Please use av_force_cpu_flags() and av_get_cpu_flags() instead which are more flexible * * @warning this function is not thread safe. */ -void av_set_cpu_flags_mask(int mask); +attribute_deprecated void av_set_cpu_flags_mask(int mask); /** * Parse CPU flags from a string. * + * The returned flags contain the specified flags as well as related unspecified flags. + * + * This function exists only for compatibility with libav. + * Please use av_parse_cpu_caps() when possible. * @return a combination of AV_CPU_* flags, negative on error. */ +attribute_deprecated int av_parse_cpu_flags(const char *s); /** + * Parse CPU caps from a string and update the given AV_CPU_* flags based on that. + * + * @return negative on error. + */ +int av_parse_cpu_caps(unsigned *flags, const char *s); + +/** * @return the number of logical CPU cores present. */ int av_cpu_count(void); diff --git a/libavutil/cpu_internal.h b/libavutil/cpu_internal.h index 49c7b070e4..6c352abe1b 100644 --- a/libavutil/cpu_internal.h +++ b/libavutil/cpu_internal.h @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -24,13 +24,16 @@ #define CPUEXT_SUFFIX(flags, suffix, cpuext) \ (HAVE_ ## cpuext ## suffix && ((flags) & AV_CPU_FLAG_ ## cpuext)) -#define CPUEXT_SUFFIX_FAST(flags, suffix, cpuext) \ +#define CPUEXT_SUFFIX_FAST2(flags, suffix, cpuext, slow_cpuext) \ (HAVE_ ## cpuext ## suffix && ((flags) & AV_CPU_FLAG_ ## cpuext) && \ - !((flags) & AV_CPU_FLAG_ ## cpuext ## SLOW)) + !((flags) & AV_CPU_FLAG_ ## slow_cpuext ## SLOW)) -#define CPUEXT_SUFFIX_SLOW(flags, suffix, cpuext) \ +#define CPUEXT_SUFFIX_SLOW2(flags, suffix, cpuext, slow_cpuext) \ (HAVE_ ## cpuext ## suffix && ((flags) & AV_CPU_FLAG_ ## cpuext) && \ - ((flags) & AV_CPU_FLAG_ ## cpuext ## SLOW)) + ((flags) & AV_CPU_FLAG_ ## slow_cpuext ## SLOW)) + +#define CPUEXT_SUFFIX_FAST(flags, suffix, cpuext) CPUEXT_SUFFIX_FAST2(flags, suffix, cpuext, cpuext) +#define CPUEXT_SUFFIX_SLOW(flags, suffix, cpuext) CPUEXT_SUFFIX_SLOW2(flags, suffix, cpuext, cpuext) #define CPUEXT(flags, cpuext) CPUEXT_SUFFIX(flags, , cpuext) #define CPUEXT_FAST(flags, cpuext) CPUEXT_SUFFIX_FAST(flags, , cpuext) diff --git a/libavutil/crc-test.c b/libavutil/crc-test.c index 41601c5ed8..3c250df919 100644 --- a/libavutil/crc-test.c +++ b/libavutil/crc-test.c @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -25,19 +25,20 @@ int main(void) { uint8_t buf[1999]; int i; - int p[5][3] = { + unsigned p[6][3] = { { AV_CRC_32_IEEE_LE, 0xEDB88320, 0x3D5CDD04 }, - { AV_CRC_32_IEEE, 0x04C11DB7, 0xC0F5BAE0 }, - { AV_CRC_16_ANSI_LE, 0xA001, 0xBFD8 }, - { AV_CRC_16_ANSI, 0x8005, 0x1FBB }, - { AV_CRC_8_ATM, 0x07, 0xE3 } + { AV_CRC_32_IEEE , 0x04C11DB7, 0xC0F5BAE0 }, + { AV_CRC_24_IEEE , 0x864CFB , 0xB704CE }, + { AV_CRC_16_ANSI_LE, 0xA001 , 0xBFD8 }, + { AV_CRC_16_ANSI , 0x8005 , 0x1FBB }, + { AV_CRC_8_ATM , 0x07 , 0xE3 } }; const AVCRC *ctx; for (i = 0; i < sizeof(buf); i++) buf[i] = i + i * i; - for (i = 0; i < 5; i++) { + for (i = 0; i < 6; i++) { ctx = av_crc_get_table(p[i][0]); printf("crc %08X = %X\n", p[i][1], av_crc(ctx, 0, buf, sizeof(buf))); } diff --git a/libavutil/crc.c b/libavutil/crc.c index 5428f8c0b6..495732b163 100644 --- a/libavutil/crc.c +++ b/libavutil/crc.c @@ -1,20 +1,20 @@ /* * copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -120,6 +120,45 @@ static const AVCRC av_crc_table[AV_CRC_MAX][257] = { 0x176E, 0x367E, 0x554E, 0x745E, 0x932E, 0xB23E, 0xD10E, 0xF01E, 0x0001 }, + [AV_CRC_24_IEEE] = { + 0x000000, 0xFB4C86, 0x0DD58A, 0xF6990C, 0xE1E693, 0x1AAA15, 0xEC3319, + 0x177F9F, 0x3981A1, 0xC2CD27, 0x34542B, 0xCF18AD, 0xD86732, 0x232BB4, + 0xD5B2B8, 0x2EFE3E, 0x894EC5, 0x720243, 0x849B4F, 0x7FD7C9, 0x68A856, + 0x93E4D0, 0x657DDC, 0x9E315A, 0xB0CF64, 0x4B83E2, 0xBD1AEE, 0x465668, + 0x5129F7, 0xAA6571, 0x5CFC7D, 0xA7B0FB, 0xE9D10C, 0x129D8A, 0xE40486, + 0x1F4800, 0x08379F, 0xF37B19, 0x05E215, 0xFEAE93, 0xD050AD, 0x2B1C2B, + 0xDD8527, 0x26C9A1, 0x31B63E, 0xCAFAB8, 0x3C63B4, 0xC72F32, 0x609FC9, + 0x9BD34F, 0x6D4A43, 0x9606C5, 0x81795A, 0x7A35DC, 0x8CACD0, 0x77E056, + 0x591E68, 0xA252EE, 0x54CBE2, 0xAF8764, 0xB8F8FB, 0x43B47D, 0xB52D71, + 0x4E61F7, 0xD2A319, 0x29EF9F, 0xDF7693, 0x243A15, 0x33458A, 0xC8090C, + 0x3E9000, 0xC5DC86, 0xEB22B8, 0x106E3E, 0xE6F732, 0x1DBBB4, 0x0AC42B, + 0xF188AD, 0x0711A1, 0xFC5D27, 0x5BEDDC, 0xA0A15A, 0x563856, 0xAD74D0, + 0xBA0B4F, 0x4147C9, 0xB7DEC5, 0x4C9243, 0x626C7D, 0x9920FB, 0x6FB9F7, + 0x94F571, 0x838AEE, 0x78C668, 0x8E5F64, 0x7513E2, 0x3B7215, 0xC03E93, + 0x36A79F, 0xCDEB19, 0xDA9486, 0x21D800, 0xD7410C, 0x2C0D8A, 0x02F3B4, + 0xF9BF32, 0x0F263E, 0xF46AB8, 0xE31527, 0x1859A1, 0xEEC0AD, 0x158C2B, + 0xB23CD0, 0x497056, 0xBFE95A, 0x44A5DC, 0x53DA43, 0xA896C5, 0x5E0FC9, + 0xA5434F, 0x8BBD71, 0x70F1F7, 0x8668FB, 0x7D247D, 0x6A5BE2, 0x911764, + 0x678E68, 0x9CC2EE, 0xA44733, 0x5F0BB5, 0xA992B9, 0x52DE3F, 0x45A1A0, + 0xBEED26, 0x48742A, 0xB338AC, 0x9DC692, 0x668A14, 0x901318, 0x6B5F9E, + 0x7C2001, 0x876C87, 0x71F58B, 0x8AB90D, 0x2D09F6, 0xD64570, 0x20DC7C, + 0xDB90FA, 0xCCEF65, 0x37A3E3, 0xC13AEF, 0x3A7669, 0x148857, 0xEFC4D1, + 0x195DDD, 0xE2115B, 0xF56EC4, 0x0E2242, 0xF8BB4E, 0x03F7C8, 0x4D963F, + 0xB6DAB9, 0x4043B5, 0xBB0F33, 0xAC70AC, 0x573C2A, 0xA1A526, 0x5AE9A0, + 0x74179E, 0x8F5B18, 0x79C214, 0x828E92, 0x95F10D, 0x6EBD8B, 0x982487, + 0x636801, 0xC4D8FA, 0x3F947C, 0xC90D70, 0x3241F6, 0x253E69, 0xDE72EF, + 0x28EBE3, 0xD3A765, 0xFD595B, 0x0615DD, 0xF08CD1, 0x0BC057, 0x1CBFC8, + 0xE7F34E, 0x116A42, 0xEA26C4, 0x76E42A, 0x8DA8AC, 0x7B31A0, 0x807D26, + 0x9702B9, 0x6C4E3F, 0x9AD733, 0x619BB5, 0x4F658B, 0xB4290D, 0x42B001, + 0xB9FC87, 0xAE8318, 0x55CF9E, 0xA35692, 0x581A14, 0xFFAAEF, 0x04E669, + 0xF27F65, 0x0933E3, 0x1E4C7C, 0xE500FA, 0x1399F6, 0xE8D570, 0xC62B4E, + 0x3D67C8, 0xCBFEC4, 0x30B242, 0x27CDDD, 0xDC815B, 0x2A1857, 0xD154D1, + 0x9F3526, 0x6479A0, 0x92E0AC, 0x69AC2A, 0x7ED3B5, 0x859F33, 0x73063F, + 0x884AB9, 0xA6B487, 0x5DF801, 0xAB610D, 0x502D8B, 0x475214, 0xBC1E92, + 0x4A879E, 0xB1CB18, 0x167BE3, 0xED3765, 0x1BAE69, 0xE0E2EF, 0xF79D70, + 0x0CD1F6, 0xFA48FA, 0x01047C, 0x2FFA42, 0xD4B6C4, 0x222FC8, 0xD9634E, + 0xCE1CD1, 0x355057, 0xC3C95B, 0x3885DD, 0x000001, + }, [AV_CRC_32_IEEE] = { 0x00000000, 0xB71DC104, 0x6E3B8209, 0xD926430D, 0xDC760413, 0x6B6BC517, 0xB24D861A, 0x0550471E, 0xB8ED0826, 0x0FF0C922, 0xD6D68A2F, 0x61CB4B2B, @@ -247,6 +286,11 @@ static const AVCRC av_crc_table[AV_CRC_MAX][257] = { }, }; #else +#if CONFIG_SMALL +#define CRC_TABLE_SIZE 257 +#else +#define CRC_TABLE_SIZE 1024 +#endif static struct { uint8_t le; uint8_t bits; @@ -255,11 +299,12 @@ static struct { [AV_CRC_8_ATM] = { 0, 8, 0x07 }, [AV_CRC_16_ANSI] = { 0, 16, 0x8005 }, [AV_CRC_16_CCITT] = { 0, 16, 0x1021 }, + [AV_CRC_24_IEEE] = { 0, 24, 0x864CFB }, [AV_CRC_32_IEEE] = { 0, 32, 0x04C11DB7 }, [AV_CRC_32_IEEE_LE] = { 1, 32, 0xEDB88320 }, [AV_CRC_16_ANSI_LE] = { 1, 16, 0xA001 }, }; -static AVCRC av_crc_table[AV_CRC_MAX][257]; +static AVCRC av_crc_table[AV_CRC_MAX][CRC_TABLE_SIZE]; #endif int av_crc_init(AVCRC *ctx, int le, int bits, uint32_t poly, int ctx_size) @@ -268,9 +313,9 @@ int av_crc_init(AVCRC *ctx, int le, int bits, uint32_t poly, int ctx_size) uint32_t c; if (bits < 8 || bits > 32 || poly >= (1LL << bits)) - return -1; + return AVERROR(EINVAL); if (ctx_size != sizeof(AVCRC) * 257 && ctx_size != sizeof(AVCRC) * 1024) - return -1; + return AVERROR(EINVAL); for (i = 0; i < 256; i++) { if (le) { diff --git a/libavutil/crc.h b/libavutil/crc.h index dcb7e2c7d7..ef8a7137e4 100644 --- a/libavutil/crc.h +++ b/libavutil/crc.h @@ -1,20 +1,20 @@ /* * copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -24,6 +24,13 @@ #include <stdint.h> #include <stddef.h> #include "attributes.h" +#include "version.h" + +/** + * @defgroup lavu_crc32 CRC32 + * @ingroup lavu_crypto + * @{ + */ typedef uint32_t AVCRC; @@ -34,6 +41,11 @@ typedef enum { AV_CRC_32_IEEE, AV_CRC_32_IEEE_LE, /*< reversed bitorder version of AV_CRC_32_IEEE */ AV_CRC_16_ANSI_LE, /*< reversed bitorder version of AV_CRC_16_ANSI */ +#if FF_API_CRC_BIG_TABLE + AV_CRC_24_IEEE = 12, +#else + AV_CRC_24_IEEE, +#endif /* FF_API_CRC_BIG_TABLE */ AV_CRC_MAX, /*< Not part of public API! Do not use outside libavutil. */ }AVCRCId; @@ -72,4 +84,8 @@ const AVCRC *av_crc_get_table(AVCRCId crc_id); uint32_t av_crc(const AVCRC *ctx, uint32_t crc, const uint8_t *buffer, size_t length) av_pure; +/** + * @} + */ + #endif /* AVUTIL_CRC_H */ diff --git a/libavutil/des-test.c b/libavutil/des-test.c index a372035a54..99453af6a7 100644 --- a/libavutil/des-test.c +++ b/libavutil/des-test.c @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -98,10 +98,10 @@ int main(void) key[1] = rand64(); key[2] = rand64(); data = rand64(); - av_des_init(&d, key, 192, 0); - av_des_crypt(&d, &ct, &data, 1, NULL, 0); - av_des_init(&d, key, 192, 1); - av_des_crypt(&d, &ct, &ct, 1, NULL, 1); + av_des_init(&d, (uint8_t *) key, 192, 0); + av_des_crypt(&d, (uint8_t *) &ct, (uint8_t *) &data, 1, NULL, 0); + av_des_init(&d, (uint8_t *) key, 192, 1); + av_des_crypt(&d, (uint8_t *) &ct, (uint8_t *) &ct, 1, NULL, 1); if (ct != data) { printf("Test 2 failed\n"); return 1; diff --git a/libavutil/des.c b/libavutil/des.c index 1dbedd92cd..d375095273 100644 --- a/libavutil/des.c +++ b/libavutil/des.c @@ -2,20 +2,20 @@ * DES encryption/decryption * Copyright (c) 2007 Reimar Doeffinger * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -27,13 +27,6 @@ #include "mem.h" #include "des.h" -#if !FF_API_CRYPTO_CONTEXT -struct AVDES { - uint64_t round_keys[3][16]; - int triple_des; -}; -#endif - #define T(a, b, c, d, e, f, g, h) 64 - a, 64 - b, 64 - c, 64 - d, 64 - e, 64 - f, 64 - g, 64 - h static const uint8_t IP_shuffle[] = { T(58, 50, 42, 34, 26, 18, 10, 2), @@ -282,10 +275,9 @@ AVDES *av_des_alloc(void) return av_mallocz(sizeof(struct AVDES)); } -int av_des_init(AVDES *d, const uint8_t *key, int key_bits, int decrypt) -{ +int av_des_init(AVDES *d, const uint8_t *key, int key_bits, av_unused int decrypt) { if (key_bits != 64 && key_bits != 192) - return -1; + return AVERROR(EINVAL); d->triple_des = key_bits > 64; gen_roundkeys(d->round_keys[0], AV_RB64(key)); if (d->triple_des) { diff --git a/libavutil/des.h b/libavutil/des.h index 1470d355ad..4cf11f5bca 100644 --- a/libavutil/des.h +++ b/libavutil/des.h @@ -2,20 +2,20 @@ * DES encryption/decryption * Copyright (c) 2007 Reimar Doeffinger * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -23,7 +23,6 @@ #define AVUTIL_DES_H #include <stdint.h> -#include "version.h" /** * @defgroup lavu_des DES @@ -31,14 +30,10 @@ * @{ */ -#if FF_API_CRYPTO_CONTEXT typedef struct AVDES { uint64_t round_keys[3][16]; int triple_des; } AVDES; -#else -typedef struct AVDES AVDES; -#endif /** * Allocate an AVDES context. diff --git a/libavutil/dict-test.c b/libavutil/dict-test.c new file mode 100644 index 0000000000..e40b57885f --- /dev/null +++ b/libavutil/dict-test.c @@ -0,0 +1,133 @@ +/* + * copyright (c) 2009 Michael Niedermayer + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "dict.c" + +static void print_dict(const AVDictionary *m) +{ + AVDictionaryEntry *t = NULL; + while ((t = av_dict_get(m, "", t, AV_DICT_IGNORE_SUFFIX))) + printf("%s %s ", t->key, t->value); + printf("\n"); +} + +static void test_separators(const AVDictionary *m, const char pair, const char val) +{ + AVDictionary *dict = NULL; + char pairs[] = {pair , '\0'}; + char vals[] = {val, '\0'}; + + char *buffer = NULL; + av_dict_copy(&dict, m, 0); + print_dict(dict); + av_dict_get_string(dict, &buffer, val, pair); + printf("%s\n", buffer); + av_dict_free(&dict); + av_dict_parse_string(&dict, buffer, vals, pairs, 0); + av_freep(&buffer); + print_dict(dict); + av_dict_free(&dict); +} + +int main(void) +{ + AVDictionary *dict = NULL; + AVDictionaryEntry *e; + char *buffer = NULL; + + printf("Testing av_dict_get_string() and av_dict_parse_string()\n"); + av_dict_get_string(dict, &buffer, '=', ','); + printf("%s\n", buffer); + av_freep(&buffer); + av_dict_set(&dict, "aaa", "aaa", 0); + av_dict_set(&dict, "b,b", "bbb", 0); + av_dict_set(&dict, "c=c", "ccc", 0); + av_dict_set(&dict, "ddd", "d,d", 0); + av_dict_set(&dict, "eee", "e=e", 0); + av_dict_set(&dict, "f,f", "f=f", 0); + av_dict_set(&dict, "g=g", "g,g", 0); + test_separators(dict, ',', '='); + av_dict_free(&dict); + av_dict_set(&dict, "aaa", "aaa", 0); + av_dict_set(&dict, "bbb", "bbb", 0); + av_dict_set(&dict, "ccc", "ccc", 0); + av_dict_set(&dict, "\\,=\'\"", "\\,=\'\"", 0); + test_separators(dict, '"', '='); + test_separators(dict, '\'', '='); + test_separators(dict, ',', '"'); + test_separators(dict, ',', '\''); + test_separators(dict, '\'', '"'); + test_separators(dict, '"', '\''); + av_dict_free(&dict); + + printf("\nTesting av_dict_set()\n"); + av_dict_set(&dict, "a", "a", 0); + av_dict_set(&dict, "b", av_strdup("b"), AV_DICT_DONT_STRDUP_VAL); + av_dict_set(&dict, av_strdup("c"), "c", AV_DICT_DONT_STRDUP_KEY); + av_dict_set(&dict, av_strdup("d"), av_strdup("d"), AV_DICT_DONT_STRDUP_KEY | AV_DICT_DONT_STRDUP_VAL); + av_dict_set(&dict, "e", "e", AV_DICT_DONT_OVERWRITE); + av_dict_set(&dict, "e", "f", AV_DICT_DONT_OVERWRITE); + av_dict_set(&dict, "f", "f", 0); + av_dict_set(&dict, "f", NULL, 0); + av_dict_set(&dict, "ff", "f", 0); + av_dict_set(&dict, "ff", "f", AV_DICT_APPEND); + e = NULL; + while ((e = av_dict_get(dict, "", e, AV_DICT_IGNORE_SUFFIX))) + printf("%s %s\n", e->key, e->value); + av_dict_free(&dict); + + av_dict_set(&dict, NULL, "a", 0); + av_dict_set(&dict, NULL, "b", 0); + av_dict_get(dict, NULL, NULL, 0); + e = NULL; + while ((e = av_dict_get(dict, "", e, AV_DICT_IGNORE_SUFFIX))) + printf("'%s' '%s'\n", e->key, e->value); + av_dict_free(&dict); + + + //valgrind sensible test + printf("\nTesting av_dict_set_int()\n"); + av_dict_set_int(&dict, "1", 1, AV_DICT_DONT_STRDUP_VAL); + av_dict_set_int(&dict, av_strdup("2"), 2, AV_DICT_DONT_STRDUP_KEY); + av_dict_set_int(&dict, av_strdup("3"), 3, AV_DICT_DONT_STRDUP_KEY | AV_DICT_DONT_STRDUP_VAL); + av_dict_set_int(&dict, "4", 4, 0); + av_dict_set_int(&dict, "5", 5, AV_DICT_DONT_OVERWRITE); + av_dict_set_int(&dict, "5", 6, AV_DICT_DONT_OVERWRITE); + av_dict_set_int(&dict, "12", 1, 0); + av_dict_set_int(&dict, "12", 2, AV_DICT_APPEND); + e = NULL; + while ((e = av_dict_get(dict, "", e, AV_DICT_IGNORE_SUFFIX))) + printf("%s %s\n", e->key, e->value); + av_dict_free(&dict); + + //valgrind sensible test + printf("\nTesting av_dict_set() with existing AVDictionaryEntry.key as key\n"); + av_dict_set(&dict, "key", "old", 0); + e = av_dict_get(dict, "key", NULL, 0); + av_dict_set(&dict, e->key, "new val OK", 0); + e = av_dict_get(dict, "key", NULL, 0); + printf("%s\n", e->value); + av_dict_set(&dict, e->key, e->value, 0); + e = av_dict_get(dict, "key", NULL, 0); + printf("%s\n", e->value); + av_dict_free(&dict); + + return 0; +} diff --git a/libavutil/dict.c b/libavutil/dict.c index 7213bf2519..f70c7e0051 100644 --- a/libavutil/dict.c +++ b/libavutil/dict.c @@ -1,20 +1,20 @@ /* * copyright (c) 2009 Michael Niedermayer * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -24,6 +24,7 @@ #include "dict.h" #include "internal.h" #include "mem.h" +#include "bprint.h" struct AVDictionary { int count; @@ -69,19 +70,29 @@ int av_dict_set(AVDictionary **pm, const char *key, const char *value, int flags) { AVDictionary *m = *pm; - AVDictionaryEntry *tag = av_dict_get(m, key, NULL, flags); - char *oldval = NULL; - int allocated = !!m; + AVDictionaryEntry *tag = NULL; + char *oldval = NULL, *copy_key = NULL, *copy_value = NULL; + if (!(flags & AV_DICT_MULTIKEY)) { + tag = av_dict_get(m, key, NULL, flags); + } + if (flags & AV_DICT_DONT_STRDUP_KEY) + copy_key = (void *)key; + else + copy_key = av_strdup(key); + if (flags & AV_DICT_DONT_STRDUP_VAL) + copy_value = (void *)value; + else if (copy_key) + copy_value = av_strdup(value); if (!m) m = *pm = av_mallocz(sizeof(*m)); - if (!m) - return AVERROR(ENOMEM); + if (!m || (key && !copy_key) || (value && !copy_value)) + goto err_out; if (tag) { if (flags & AV_DICT_DONT_OVERWRITE) { - if (flags & AV_DICT_DONT_STRDUP_KEY) av_free(key); - if (flags & AV_DICT_DONT_STRDUP_VAL) av_free(value); + av_free(copy_key); + av_free(copy_value); return 0; } if (flags & AV_DICT_APPEND) @@ -90,39 +101,55 @@ int av_dict_set(AVDictionary **pm, const char *key, const char *value, av_free(tag->value); av_free(tag->key); *tag = m->elems[--m->count]; - } else { - int ret = av_reallocp_array(&m->elems, - m->count + 1, sizeof(*m->elems)); - if (ret < 0) { - if (allocated) - av_freep(pm); - - return ret; - } + } else if (copy_value) { + AVDictionaryEntry *tmp = av_realloc(m->elems, + (m->count + 1) * sizeof(*m->elems)); + if (!tmp) + goto err_out; + m->elems = tmp; } - if (value) { - if (flags & AV_DICT_DONT_STRDUP_KEY) - m->elems[m->count].key = key; - else - m->elems[m->count].key = av_strdup(key); - if (flags & AV_DICT_DONT_STRDUP_VAL) { - m->elems[m->count].value = value; - } else if (oldval && flags & AV_DICT_APPEND) { - int len = strlen(oldval) + strlen(value) + 1; - if (!(oldval = av_realloc(oldval, len))) - return AVERROR(ENOMEM); - av_strlcat(oldval, value, len); - m->elems[m->count].value = oldval; - } else - m->elems[m->count].value = av_strdup(value); + if (copy_value) { + m->elems[m->count].key = copy_key; + m->elems[m->count].value = copy_value; + if (oldval && flags & AV_DICT_APPEND) { + size_t len = strlen(oldval) + strlen(copy_value) + 1; + char *newval = av_mallocz(len); + if (!newval) + goto err_out; + av_strlcat(newval, oldval, len); + av_freep(&oldval); + av_strlcat(newval, copy_value, len); + m->elems[m->count].value = newval; + av_freep(©_value); + } m->count++; + } else { + av_freep(©_key); } if (!m->count) { - av_free(m->elems); + av_freep(&m->elems); av_freep(pm); } return 0; + +err_out: + if (m && !m->count) { + av_freep(&m->elems); + av_freep(pm); + } + av_free(copy_key); + av_free(copy_value); + return AVERROR(ENOMEM); +} + +int av_dict_set_int(AVDictionary **pm, const char *key, int64_t value, + int flags) +{ + char valuestr[22]; + snprintf(valuestr, sizeof(valuestr), "%"PRId64, value); + flags &= ~AV_DICT_DONT_STRDUP_VAL; + return av_dict_set(pm, key, valuestr, flags); } static int parse_key_value_pair(AVDictionary **pm, const char **buf, @@ -178,10 +205,10 @@ void av_dict_free(AVDictionary **pm) if (m) { while (m->count--) { - av_free(m->elems[m->count].key); - av_free(m->elems[m->count].value); + av_freep(&m->elems[m->count].key); + av_freep(&m->elems[m->count].value); } - av_free(m->elems); + av_freep(&m->elems); } av_freep(pm); } @@ -198,3 +225,31 @@ int av_dict_copy(AVDictionary **dst, const AVDictionary *src, int flags) return 0; } + +int av_dict_get_string(const AVDictionary *m, char **buffer, + const char key_val_sep, const char pairs_sep) +{ + AVDictionaryEntry *t = NULL; + AVBPrint bprint; + int cnt = 0; + char special_chars[] = {pairs_sep, key_val_sep, '\0'}; + + if (!buffer || pairs_sep == '\0' || key_val_sep == '\0' || pairs_sep == key_val_sep || + pairs_sep == '\\' || key_val_sep == '\\') + return AVERROR(EINVAL); + + if (!av_dict_count(m)) { + *buffer = av_strdup(""); + return *buffer ? 0 : AVERROR(ENOMEM); + } + + av_bprint_init(&bprint, 64, AV_BPRINT_SIZE_UNLIMITED); + while ((t = av_dict_get(m, "", t, AV_DICT_IGNORE_SUFFIX))) { + if (cnt++) + av_bprint_append_data(&bprint, &pairs_sep, 1); + av_bprint_escape(&bprint, t->key, special_chars, AV_ESCAPE_MODE_BACKSLASH, 0); + av_bprint_append_data(&bprint, &key_val_sep, 1); + av_bprint_escape(&bprint, t->value, special_chars, AV_ESCAPE_MODE_BACKSLASH, 0); + } + return av_bprint_finalize(&bprint, buffer); +} diff --git a/libavutil/dict.h b/libavutil/dict.h index 3649b9fbe7..118f1f00ed 100644 --- a/libavutil/dict.h +++ b/libavutil/dict.h @@ -1,29 +1,39 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ /** * @file * Public dictionary API. + * @deprecated + * AVDictionary is provided for compatibility with libav. It is both in + * implementation as well as API inefficient. It does not scale and is + * extremely slow with large dictionaries. + * It is recommended that new code uses our tree container from tree.c/h + * where applicable, which uses AVL trees to achieve O(log n) performance. */ #ifndef AVUTIL_DICT_H #define AVUTIL_DICT_H +#include <stdint.h> + +#include "version.h" + /** * @addtogroup lavu_dict AVDictionary * @ingroup lavu_data @@ -56,15 +66,17 @@ @endcode */ -#define AV_DICT_MATCH_CASE 1 -#define AV_DICT_IGNORE_SUFFIX 2 +#define AV_DICT_MATCH_CASE 1 /**< Only get an entry with exact-case key match. Only relevant in av_dict_get(). */ +#define AV_DICT_IGNORE_SUFFIX 2 /**< Return first entry in a dictionary whose first part corresponds to the search key, + ignoring the suffix of the found key string. Only relevant in av_dict_get(). */ #define AV_DICT_DONT_STRDUP_KEY 4 /**< Take ownership of a key that's been - allocated with av_malloc() and children. */ + allocated with av_malloc() or another memory allocation function. */ #define AV_DICT_DONT_STRDUP_VAL 8 /**< Take ownership of a value that's been - allocated with av_malloc() and chilren. */ + allocated with av_malloc() or another memory allocation function. */ #define AV_DICT_DONT_OVERWRITE 16 ///< Don't overwrite existing entries. #define AV_DICT_APPEND 32 /**< If the entry already exists, append to it. Note that no delimiter is added, the strings are simply concatenated. */ +#define AV_DICT_MULTIKEY 64 /**< Allow to store several equal keys in the dictionary */ typedef struct AVDictionaryEntry { char *key; @@ -76,10 +88,17 @@ typedef struct AVDictionary AVDictionary; /** * Get a dictionary entry with matching key. * + * The returned entry key or value must not be changed, or it will + * cause undefined behavior. + * + * To iterate through all the dictionary entries, you can set the matching key + * to the null string "" and set the AV_DICT_IGNORE_SUFFIX flag. + * * @param prev Set to the previous matching element to find the next. * If set to NULL the first matching element is returned. - * @param flags Allows case as well as suffix-insensitive comparisons. - * @return Found entry or NULL, changing key or value leads to undefined behavior. + * @param key matching key + * @param flags a collection of AV_DICT_* flags controlling how the entry is retrieved + * @return found entry or NULL in case no matching entry was found in the dictionary */ AVDictionaryEntry *av_dict_get(const AVDictionary *m, const char *key, const AVDictionaryEntry *prev, int flags); @@ -95,17 +114,34 @@ int av_dict_count(const AVDictionary *m); /** * Set the given entry in *pm, overwriting an existing entry. * + * Note: If AV_DICT_DONT_STRDUP_KEY or AV_DICT_DONT_STRDUP_VAL is set, + * these arguments will be freed on error. + * + * Warning: Adding a new entry to a dictionary invalidates all existing entries + * previously returned with av_dict_get. + * * @param pm pointer to a pointer to a dictionary struct. If *pm is NULL * a dictionary struct is allocated and put in *pm. - * @param key entry key to add to *pm (will be av_strduped depending on flags) - * @param value entry value to add to *pm (will be av_strduped depending on flags). + * @param key entry key to add to *pm (will either be av_strduped or added as a new key depending on flags) + * @param value entry value to add to *pm (will be av_strduped or added as a new key depending on flags). * Passing a NULL value will cause an existing entry to be deleted. * @return >= 0 on success otherwise an error code <0 */ int av_dict_set(AVDictionary **pm, const char *key, const char *value, int flags); /** - * Parse the key/value pairs list and add to a dictionary. + * Convenience wrapper for av_dict_set that converts the value to a string + * and stores it. + * + * Note: If AV_DICT_DONT_STRDUP_KEY is set, key will be freed on error. + */ +int av_dict_set_int(AVDictionary **pm, const char *key, int64_t value, int flags); + +/** + * Parse the key/value pairs list and add the parsed entries to a dictionary. + * + * In case of failure, all the successfully set entries are stored in + * *pm. You may need to manually free the created dictionary. * * @param key_val_sep a 0-terminated list of characters used to separate * key from value @@ -140,6 +176,24 @@ int av_dict_copy(AVDictionary **dst, const AVDictionary *src, int flags); void av_dict_free(AVDictionary **m); /** + * Get dictionary entries as a string. + * + * Create a string containing dictionary's entries. + * Such string may be passed back to av_dict_parse_string(). + * @note String is escaped with backslashes ('\'). + * + * @param[in] m dictionary + * @param[out] buffer Pointer to buffer that will be allocated with string containg entries. + * Buffer must be freed by the caller when is no longer needed. + * @param[in] key_val_sep character used to separate key from value + * @param[in] pairs_sep character used to separate two pairs from each other + * @return >= 0 on success, negative on error + * @warning Separators cannot be neither '\\' nor '\0'. They also cannot be the same. + */ +int av_dict_get_string(const AVDictionary *m, char **buffer, + const char key_val_sep, const char pairs_sep); + +/** * @} */ diff --git a/libavutil/display-test.c b/libavutil/display-test.c new file mode 100644 index 0000000000..b42ab893a8 --- /dev/null +++ b/libavutil/display-test.c @@ -0,0 +1,61 @@ +/* + * Copyright (c) 2014 Vittorio Giovara <vittorio.giovara@gmail.com> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "display.c" + +static void print_matrix(int32_t matrix[9]) +{ + int i, j; + + for (i = 0; i < 3; ++i) { + for (j = 0; j < 3 - 1; ++j) + printf("%d ", matrix[i*3 + j]); + + printf("%d\n", matrix[i*3 + j]); + } +} + +int main(void) +{ + int32_t matrix[9]; + + // Set the matrix to 90 degrees + av_display_rotation_set(matrix, 90); + print_matrix(matrix); + printf("degrees: %f\n", av_display_rotation_get(matrix)); + + // Set the matrix to -45 degrees + av_display_rotation_set(matrix, -45); + print_matrix(matrix); + printf("degrees: %f\n", av_display_rotation_get(matrix)); + + // flip horizontal + av_display_matrix_flip(matrix, 1, 0); + print_matrix(matrix); + printf("degrees: %f\n", av_display_rotation_get(matrix)); + + // flip vertical + av_display_matrix_flip(matrix, 0, 1); + print_matrix(matrix); + printf("degrees: %f\n", av_display_rotation_get(matrix)); + + return 0; + +} diff --git a/libavutil/display.c b/libavutil/display.c index f7500948ff..a0076e067b 100644 --- a/libavutil/display.c +++ b/libavutil/display.c @@ -1,20 +1,20 @@ /* * Copyright (c) 2014 Vittorio Giovara <vittorio.giovara@gmail.com> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -35,10 +35,8 @@ double av_display_rotation_get(const int32_t matrix[9]) { double rotation, scale[2]; - scale[0] = sqrt(CONV_FP(matrix[0]) * CONV_FP(matrix[0]) + - CONV_FP(matrix[3]) * CONV_FP(matrix[3])); - scale[1] = sqrt(CONV_FP(matrix[1]) * CONV_FP(matrix[1]) + - CONV_FP(matrix[4]) * CONV_FP(matrix[4])); + scale[0] = hypot(CONV_FP(matrix[0]), CONV_FP(matrix[3])); + scale[1] = hypot(CONV_FP(matrix[1]), CONV_FP(matrix[4])); if (scale[0] == 0.0 || scale[1] == 0.0) return NAN; diff --git a/libavutil/display.h b/libavutil/display.h index dba3b1e60d..39c15ee6b8 100644 --- a/libavutil/display.h +++ b/libavutil/display.h @@ -1,20 +1,20 @@ /* * Copyright (c) 2014 Vittorio Giovara <vittorio.giovara@gmail.com> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -22,6 +22,7 @@ #define AVUTIL_DISPLAY_H #include <stdint.h> +#include "common.h" /** * The display transformation matrix specifies an affine transformation that diff --git a/libavutil/downmix_info.c b/libavutil/downmix_info.c index 51505c8e04..c634c6a79f 100644 --- a/libavutil/downmix_info.c +++ b/libavutil/downmix_info.c @@ -1,20 +1,20 @@ /* * Copyright (c) 2014 Tim Walker <tdskywalker@gmail.com> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/downmix_info.h b/libavutil/downmix_info.h index 69969f6fbd..221cf5bf9b 100644 --- a/libavutil/downmix_info.h +++ b/libavutil/downmix_info.h @@ -1,20 +1,20 @@ /* * Copyright (c) 2014 Tim Walker <tdskywalker@gmail.com> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -95,11 +95,12 @@ typedef struct AVDownmixInfo { /** * Get a frame's AV_FRAME_DATA_DOWNMIX_INFO side data for editing. * - * The side data is created and added to the frame if it's absent. + * If the side data is absent, it is created and added to the frame. * - * @param frame the frame for which the side data is to be obtained. + * @param frame the frame for which the side data is to be obtained or created * - * @return the AVDownmixInfo structure to be edited by the caller. + * @return the AVDownmixInfo structure to be edited by the caller, or NULL if + * the structure cannot be allocated. */ AVDownmixInfo *av_downmix_info_update_side_data(AVFrame *frame); diff --git a/libavutil/dynarray.h b/libavutil/dynarray.h new file mode 100644 index 0000000000..4947d93dc8 --- /dev/null +++ b/libavutil/dynarray.h @@ -0,0 +1,70 @@ +/* + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public License + * as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public License + * along with FFmpeg; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#ifndef AVUTIL_DYNARRAY_H +#define AVUTIL_DYNARRAY_H + +#include "log.h" +#include "mem.h" + +/** + * Add an element of to a dynamic array. + * + * The array is reallocated when its number of elements reaches powers of 2. + * Therefore, the amortized cost of adding an element is constant. + * + * In case of success, the pointer to the array is updated in order to + * point to the new grown array, and the size is incremented. + * + * @param av_size_max maximum size of the array, usually the MAX macro of + * the type of the size + * @param av_elt_size size of the elements in the array, in bytes + * @param av_array pointer to the array, must be a lvalue + * @param av_size size of the array, must be an integer lvalue + * @param av_success statement to execute on success; at this point, the + * size variable is not yet incremented + * @param av_failure statement to execute on failure; if this happens, the + * array and size are not changed; the statement can end + * with a return or a goto + */ +#define AV_DYNARRAY_ADD(av_size_max, av_elt_size, av_array, av_size, \ + av_success, av_failure) \ + do { \ + size_t av_size_new = (av_size); \ + if (!((av_size) & ((av_size) - 1))) { \ + av_size_new = (av_size) ? (av_size) << 1 : 1; \ + if (av_size_new > (av_size_max) / (av_elt_size)) { \ + av_size_new = 0; \ + } else { \ + void *av_array_new = \ + av_realloc((av_array), av_size_new * (av_elt_size)); \ + if (!av_array_new) \ + av_size_new = 0; \ + else \ + (av_array) = av_array_new; \ + } \ + } \ + if (av_size_new) { \ + { av_success } \ + (av_size)++; \ + } else { \ + av_failure \ + } \ + } while (0) + +#endif /* AVUTIL_DYNARRAY_H */ diff --git a/libavutil/error-test.c b/libavutil/error-test.c new file mode 100644 index 0000000000..36240daa4e --- /dev/null +++ b/libavutil/error-test.c @@ -0,0 +1,37 @@ +/* + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "error.c" + +#undef printf + +int main(void) +{ + int i; + + for (i = 0; i < FF_ARRAY_ELEMS(error_entries); i++) { + const struct error_entry *entry = &error_entries[i]; + printf("%d: %s [%s]\n", entry->num, av_err2str(entry->num), entry->tag); + } + + for (i = 0; i < 256; i++) { + printf("%d: %s\n", -i, av_err2str(-i)); + } + + return 0; +} diff --git a/libavutil/error.c b/libavutil/error.c index 6803d2d4cd..b96304837b 100644 --- a/libavutil/error.c +++ b/libavutil/error.c @@ -1,54 +1,123 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ +#undef _GNU_SOURCE +#define _XOPEN_SOURCE 600 /* XSI-compliant version of strerror_r */ #include "avutil.h" #include "avstring.h" #include "common.h" +struct error_entry { + int num; + const char *tag; + const char *str; +}; + +#define ERROR_TAG(tag) AVERROR_##tag, #tag +#define EERROR_TAG(tag) AVERROR(tag), #tag +#define AVERROR_INPUT_AND_OUTPUT_CHANGED (AVERROR_INPUT_CHANGED | AVERROR_OUTPUT_CHANGED) +static const struct error_entry error_entries[] = { + { ERROR_TAG(BSF_NOT_FOUND), "Bitstream filter not found" }, + { ERROR_TAG(BUG), "Internal bug, should not have happened" }, + { ERROR_TAG(BUG2), "Internal bug, should not have happened" }, + { ERROR_TAG(BUFFER_TOO_SMALL), "Buffer too small" }, + { ERROR_TAG(DECODER_NOT_FOUND), "Decoder not found" }, + { ERROR_TAG(DEMUXER_NOT_FOUND), "Demuxer not found" }, + { ERROR_TAG(ENCODER_NOT_FOUND), "Encoder not found" }, + { ERROR_TAG(EOF), "End of file" }, + { ERROR_TAG(EXIT), "Immediate exit requested" }, + { ERROR_TAG(EXTERNAL), "Generic error in an external library" }, + { ERROR_TAG(FILTER_NOT_FOUND), "Filter not found" }, + { ERROR_TAG(INPUT_CHANGED), "Input changed" }, + { ERROR_TAG(INVALIDDATA), "Invalid data found when processing input" }, + { ERROR_TAG(MUXER_NOT_FOUND), "Muxer not found" }, + { ERROR_TAG(OPTION_NOT_FOUND), "Option not found" }, + { ERROR_TAG(OUTPUT_CHANGED), "Output changed" }, + { ERROR_TAG(PATCHWELCOME), "Not yet implemented in FFmpeg, patches welcome" }, + { ERROR_TAG(PROTOCOL_NOT_FOUND), "Protocol not found" }, + { ERROR_TAG(STREAM_NOT_FOUND), "Stream not found" }, + { ERROR_TAG(UNKNOWN), "Unknown error occurred" }, + { ERROR_TAG(EXPERIMENTAL), "Experimental feature" }, + { ERROR_TAG(INPUT_AND_OUTPUT_CHANGED), "Input and output changed" }, + { ERROR_TAG(HTTP_BAD_REQUEST), "Server returned 400 Bad Request" }, + { ERROR_TAG(HTTP_UNAUTHORIZED), "Server returned 401 Unauthorized (authorization failed)" }, + { ERROR_TAG(HTTP_FORBIDDEN), "Server returned 403 Forbidden (access denied)" }, + { ERROR_TAG(HTTP_NOT_FOUND), "Server returned 404 Not Found" }, + { ERROR_TAG(HTTP_OTHER_4XX), "Server returned 4XX Client Error, but not one of 40{0,1,3,4}" }, + { ERROR_TAG(HTTP_SERVER_ERROR), "Server returned 5XX Server Error reply" }, +#if !HAVE_STRERROR_R + { EERROR_TAG(E2BIG), "Argument list too long" }, + { EERROR_TAG(EACCES), "Permission denied" }, + { EERROR_TAG(EAGAIN), "Resource temporarily unavailable" }, + { EERROR_TAG(EBADF), "Bad file descriptor" }, + { EERROR_TAG(EBUSY), "Device or resource busy" }, + { EERROR_TAG(ECHILD), "No child processes" }, + { EERROR_TAG(EDEADLK), "Resource deadlock avoided" }, + { EERROR_TAG(EDOM), "Numerical argument out of domain" }, + { EERROR_TAG(EEXIST), "File exists" }, + { EERROR_TAG(EFAULT), "Bad address" }, + { EERROR_TAG(EFBIG), "File too large" }, + { EERROR_TAG(EILSEQ), "Illegal byte sequence" }, + { EERROR_TAG(EINTR), "Interrupted system call" }, + { EERROR_TAG(EINVAL), "Invalid argument" }, + { EERROR_TAG(EIO), "I/O error" }, + { EERROR_TAG(EISDIR), "Is a directory" }, + { EERROR_TAG(EMFILE), "Too many open files" }, + { EERROR_TAG(EMLINK), "Too many links" }, + { EERROR_TAG(ENAMETOOLONG), "File name too long" }, + { EERROR_TAG(ENFILE), "Too many open files in system" }, + { EERROR_TAG(ENODEV), "No such device" }, + { EERROR_TAG(ENOENT), "No such file or directory" }, + { EERROR_TAG(ENOEXEC), "Exec format error" }, + { EERROR_TAG(ENOLCK), "No locks available" }, + { EERROR_TAG(ENOMEM), "Cannot allocate memory" }, + { EERROR_TAG(ENOSPC), "No space left on device" }, + { EERROR_TAG(ENOSYS), "Function not implemented" }, + { EERROR_TAG(ENOTDIR), "Not a directory" }, + { EERROR_TAG(ENOTEMPTY), "Directory not empty" }, + { EERROR_TAG(ENOTTY), "Inappropriate I/O control operation" }, + { EERROR_TAG(ENXIO), "No such device or address" }, + { EERROR_TAG(EPERM), "Operation not permitted" }, + { EERROR_TAG(EPIPE), "Broken pipe" }, + { EERROR_TAG(ERANGE), "Result too large" }, + { EERROR_TAG(EROFS), "Read-only file system" }, + { EERROR_TAG(ESPIPE), "Illegal seek" }, + { EERROR_TAG(ESRCH), "No such process" }, + { EERROR_TAG(EXDEV), "Cross-device link" }, +#endif +}; + int av_strerror(int errnum, char *errbuf, size_t errbuf_size) { - int ret = 0; - const char *errstr = NULL; + int ret = 0, i; + const struct error_entry *entry = NULL; - switch (errnum) { - case AVERROR_BSF_NOT_FOUND: errstr = "Bitstream filter not found" ; break; - case AVERROR_DECODER_NOT_FOUND: errstr = "Decoder not found" ; break; - case AVERROR_DEMUXER_NOT_FOUND: errstr = "Demuxer not found" ; break; - case AVERROR_ENCODER_NOT_FOUND: errstr = "Encoder not found" ; break; - case AVERROR_EOF: errstr = "End of file" ; break; - case AVERROR_EXIT: errstr = "Immediate exit requested" ; break; - case AVERROR_FILTER_NOT_FOUND: errstr = "Filter not found" ; break; - case AVERROR_INVALIDDATA: errstr = "Invalid data found when processing input" ; break; - case AVERROR_MUXER_NOT_FOUND: errstr = "Muxer not found" ; break; - case AVERROR_OPTION_NOT_FOUND: errstr = "Option not found" ; break; - case AVERROR_PATCHWELCOME: errstr = "Not yet implemented in Libav, patches welcome"; break; - case AVERROR_PROTOCOL_NOT_FOUND:errstr = "Protocol not found" ; break; - case AVERROR_STREAM_NOT_FOUND: errstr = "Stream not found" ; break; - case AVERROR_BUG: errstr = "Bug detected, please report the issue" ; break; - case AVERROR_UNKNOWN: errstr = "Unknown error occurred" ; break; - case AVERROR_EXPERIMENTAL: errstr = "Experimental feature" ; break; + for (i = 0; i < FF_ARRAY_ELEMS(error_entries); i++) { + if (errnum == error_entries[i].num) { + entry = &error_entries[i]; + break; + } } - - if (errstr) { - av_strlcpy(errbuf, errstr, errbuf_size); + if (entry) { + av_strlcpy(errbuf, entry->str, errbuf_size); } else { #if HAVE_STRERROR_R - ret = strerror_r(AVUNERROR(errnum), errbuf, errbuf_size); + ret = AVERROR(strerror_r(AVUNERROR(errnum), errbuf, errbuf_size)); #else ret = -1; #endif diff --git a/libavutil/error.h b/libavutil/error.h index 8be87cbeba..71df4da353 100644 --- a/libavutil/error.h +++ b/libavutil/error.h @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -44,24 +44,43 @@ #define AVUNERROR(e) (e) #endif -#define AVERROR_BSF_NOT_FOUND (-0x39acbd08) ///< Bitstream filter not found -#define AVERROR_DECODER_NOT_FOUND (-0x3cbabb08) ///< Decoder not found -#define AVERROR_DEMUXER_NOT_FOUND (-0x32babb08) ///< Demuxer not found -#define AVERROR_ENCODER_NOT_FOUND (-0x3cb1ba08) ///< Encoder not found -#define AVERROR_EOF (-0x5fb9b0bb) ///< End of file -#define AVERROR_EXIT (-0x2bb6a7bb) ///< Immediate exit was requested; the called function should not be restarted -#define AVERROR_FILTER_NOT_FOUND (-0x33b6b908) ///< Filter not found -#define AVERROR_INVALIDDATA (-0x3ebbb1b7) ///< Invalid data found when processing input -#define AVERROR_MUXER_NOT_FOUND (-0x27aab208) ///< Muxer not found -#define AVERROR_OPTION_NOT_FOUND (-0x2bafb008) ///< Option not found -#define AVERROR_PATCHWELCOME (-0x3aa8beb0) ///< Not yet implemented in Libav, patches welcome -#define AVERROR_PROTOCOL_NOT_FOUND (-0x30adaf08) ///< Protocol not found -#define AVERROR_STREAM_NOT_FOUND (-0x2dabac08) ///< Stream not found -#define AVERROR_BUG (-0x5fb8aabe) ///< Bug detected, please report the issue -#define AVERROR_UNKNOWN (-0x31b4b1ab) ///< Unknown error, typically from an external library +#define FFERRTAG(a, b, c, d) (-(int)MKTAG(a, b, c, d)) + +#define AVERROR_BSF_NOT_FOUND FFERRTAG(0xF8,'B','S','F') ///< Bitstream filter not found +#define AVERROR_BUG FFERRTAG( 'B','U','G','!') ///< Internal bug, also see AVERROR_BUG2 +#define AVERROR_BUFFER_TOO_SMALL FFERRTAG( 'B','U','F','S') ///< Buffer too small +#define AVERROR_DECODER_NOT_FOUND FFERRTAG(0xF8,'D','E','C') ///< Decoder not found +#define AVERROR_DEMUXER_NOT_FOUND FFERRTAG(0xF8,'D','E','M') ///< Demuxer not found +#define AVERROR_ENCODER_NOT_FOUND FFERRTAG(0xF8,'E','N','C') ///< Encoder not found +#define AVERROR_EOF FFERRTAG( 'E','O','F',' ') ///< End of file +#define AVERROR_EXIT FFERRTAG( 'E','X','I','T') ///< Immediate exit was requested; the called function should not be restarted +#define AVERROR_EXTERNAL FFERRTAG( 'E','X','T',' ') ///< Generic error in an external library +#define AVERROR_FILTER_NOT_FOUND FFERRTAG(0xF8,'F','I','L') ///< Filter not found +#define AVERROR_INVALIDDATA FFERRTAG( 'I','N','D','A') ///< Invalid data found when processing input +#define AVERROR_MUXER_NOT_FOUND FFERRTAG(0xF8,'M','U','X') ///< Muxer not found +#define AVERROR_OPTION_NOT_FOUND FFERRTAG(0xF8,'O','P','T') ///< Option not found +#define AVERROR_PATCHWELCOME FFERRTAG( 'P','A','W','E') ///< Not yet implemented in FFmpeg, patches welcome +#define AVERROR_PROTOCOL_NOT_FOUND FFERRTAG(0xF8,'P','R','O') ///< Protocol not found + +#define AVERROR_STREAM_NOT_FOUND FFERRTAG(0xF8,'S','T','R') ///< Stream not found +/** + * This is semantically identical to AVERROR_BUG + * it has been introduced in Libav after our AVERROR_BUG and with a modified value. + */ +#define AVERROR_BUG2 FFERRTAG( 'B','U','G',' ') +#define AVERROR_UNKNOWN FFERRTAG( 'U','N','K','N') ///< Unknown error, typically from an external library #define AVERROR_EXPERIMENTAL (-0x2bb2afa8) ///< Requested feature is flagged experimental. Set strict_std_compliance if you really want to use it. #define AVERROR_INPUT_CHANGED (-0x636e6701) ///< Input changed between calls. Reconfiguration is required. (can be OR-ed with AVERROR_OUTPUT_CHANGED) #define AVERROR_OUTPUT_CHANGED (-0x636e6702) ///< Output changed between calls. Reconfiguration is required. (can be OR-ed with AVERROR_INPUT_CHANGED) +/* HTTP & RTSP errors */ +#define AVERROR_HTTP_BAD_REQUEST FFERRTAG(0xF8,'4','0','0') +#define AVERROR_HTTP_UNAUTHORIZED FFERRTAG(0xF8,'4','0','1') +#define AVERROR_HTTP_FORBIDDEN FFERRTAG(0xF8,'4','0','3') +#define AVERROR_HTTP_NOT_FOUND FFERRTAG(0xF8,'4','0','4') +#define AVERROR_HTTP_OTHER_4XX FFERRTAG(0xF8,'4','X','X') +#define AVERROR_HTTP_SERVER_ERROR FFERRTAG(0xF8,'5','X','X') + +#define AV_ERROR_MAX_STRING_SIZE 64 /** * Put a description of the AVERROR code errnum in errbuf. @@ -78,6 +97,29 @@ int av_strerror(int errnum, char *errbuf, size_t errbuf_size); /** + * Fill the provided buffer with a string containing an error string + * corresponding to the AVERROR code errnum. + * + * @param errbuf a buffer + * @param errbuf_size size in bytes of errbuf + * @param errnum error code to describe + * @return the buffer in input, filled with the error description + * @see av_strerror() + */ +static inline char *av_make_error_string(char *errbuf, size_t errbuf_size, int errnum) +{ + av_strerror(errnum, errbuf, errbuf_size); + return errbuf; +} + +/** + * Convenience macro, the return value should be used only directly in + * function arguments but never stand-alone. + */ +#define av_err2str(errnum) \ + av_make_error_string((char[AV_ERROR_MAX_STRING_SIZE]){0}, AV_ERROR_MAX_STRING_SIZE, errnum) + +/** * @} */ diff --git a/libavutil/eval-test.c b/libavutil/eval-test.c index da8721de7c..7646e4db61 100644 --- a/libavutil/eval-test.c +++ b/libavutil/eval-test.c @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -107,6 +107,35 @@ int main(int argc, char **argv) "not(0)", "6.0206dB", "-3.0103dB", + "pow(0,1.23)", + "pow(PI,1.23)", + "PI^1.23", + "pow(-1,1.23)", + "if(1, 2)", + "if(1, 1, 2)", + "if(0, 1, 2)", + "ifnot(0, 23)", + "ifnot(1, NaN) + if(0, 1)", + "ifnot(1, 1, 2)", + "ifnot(0, 1, 2)", + "taylor(1, 1)", + "taylor(eq(mod(ld(1),4),1)-eq(mod(ld(1),4),3), PI/2, 1)", + "root(sin(ld(0))-1, 2)", + "root(sin(ld(0))+6+sin(ld(0)/12)-log(ld(0)), 100)", + "7000000B*random(0)", + "squish(2)", + "gauss(0.1)", + "hypot(4,3)", + "gcd(30,55)*print(min(9,1))", + "bitor(42, 12)", + "bitand(42, 12)", + "bitand(NAN, 1)", + "between(10, -3, 10)", + "between(-4, -2, -1)", + "between(1,2)", + "clip(0, 2, 1)", + "clip(0/0, 1, 2)", + "clip(0, 0/0, 1)", NULL }; diff --git a/libavutil/eval.c b/libavutil/eval.c index b5d71c6000..d291c5d6d5 100644 --- a/libavutil/eval.c +++ b/libavutil/eval.c @@ -2,20 +2,20 @@ * Copyright (c) 2002-2006 Michael Niedermayer <michaelni@gmx.at> * Copyright (c) 2006 Oded Shimon <ods15@ods15.dyndns.org> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -26,12 +26,16 @@ * see http://joe.hotchkiss.com/programming/eval/eval.html */ +#include <float.h> #include "attributes.h" #include "avutil.h" #include "common.h" #include "eval.h" +#include "ffmath.h" +#include "internal.h" #include "log.h" #include "mathematics.h" +#include "time.h" #include "avstring.h" #include "timer.h" @@ -49,53 +53,70 @@ typedef struct Parser { int log_offset; void *log_ctx; #define VARS 10 - double var[VARS]; + double *var; } Parser; -static const AVClass class = { "Eval", av_default_item_name, NULL, LIBAVUTIL_VERSION_INT, offsetof(Parser,log_offset), offsetof(Parser,log_ctx) }; - -static const int8_t si_prefixes['z' - 'E' + 1] = { - ['y'-'E']= -24, - ['z'-'E']= -21, - ['a'-'E']= -18, - ['f'-'E']= -15, - ['p'-'E']= -12, - ['n'-'E']= - 9, - ['u'-'E']= - 6, - ['m'-'E']= - 3, - ['c'-'E']= - 2, - ['d'-'E']= - 1, - ['h'-'E']= 2, - ['k'-'E']= 3, - ['K'-'E']= 3, - ['M'-'E']= 6, - ['G'-'E']= 9, - ['T'-'E']= 12, - ['P'-'E']= 15, - ['E'-'E']= 18, - ['Z'-'E']= 21, - ['Y'-'E']= 24, +static const AVClass eval_class = { "Eval", av_default_item_name, NULL, LIBAVUTIL_VERSION_INT, offsetof(Parser,log_offset), offsetof(Parser,log_ctx) }; + +static const struct { + double bin_val; + double dec_val; + int8_t exp; +} si_prefixes['z' - 'E' + 1] = { + ['y'-'E']= { 8.271806125530276749e-25, 1e-24, -24 }, + ['z'-'E']= { 8.4703294725430034e-22, 1e-21, -21 }, + ['a'-'E']= { 8.6736173798840355e-19, 1e-18, -18 }, + ['f'-'E']= { 8.8817841970012523e-16, 1e-15, -15 }, + ['p'-'E']= { 9.0949470177292824e-13, 1e-12, -12 }, + ['n'-'E']= { 9.3132257461547852e-10, 1e-9, -9 }, + ['u'-'E']= { 9.5367431640625e-7, 1e-6, -6 }, + ['m'-'E']= { 9.765625e-4, 1e-3, -3 }, + ['c'-'E']= { 9.8431332023036951e-3, 1e-2, -2 }, + ['d'-'E']= { 9.921256574801246e-2, 1e-1, -1 }, + ['h'-'E']= { 1.0159366732596479e2, 1e2, 2 }, + ['k'-'E']= { 1.024e3, 1e3, 3 }, + ['K'-'E']= { 1.024e3, 1e3, 3 }, + ['M'-'E']= { 1.048576e6, 1e6, 6 }, + ['G'-'E']= { 1.073741824e9, 1e9, 9 }, + ['T'-'E']= { 1.099511627776e12, 1e12, 12 }, + ['P'-'E']= { 1.125899906842624e15, 1e15, 15 }, + ['E'-'E']= { 1.152921504606847e18, 1e18, 18 }, + ['Z'-'E']= { 1.1805916207174113e21, 1e21, 21 }, + ['Y'-'E']= { 1.2089258196146292e24, 1e24, 24 }, +}; + +static const struct { + const char *name; + double value; +} constants[] = { + { "E", M_E }, + { "PI", M_PI }, + { "PHI", M_PHI }, + { "QP2LAMBDA", FF_QP2LAMBDA }, }; double av_strtod(const char *numstr, char **tail) { double d; char *next; - d = strtod(numstr, &next); + if(numstr[0]=='0' && (numstr[1]|0x20)=='x') { + d = strtoul(numstr, &next, 16); + } else + d = strtod(numstr, &next); /* if parsing succeeded, check for and interpret postfixes */ if (next!=numstr) { if (next[0] == 'd' && next[1] == 'B') { /* treat dB as decibels instead of decibytes */ - d = pow(10, d / 20); + d = ff_exp10(d / 20); next += 2; } else if (*next >= 'E' && *next <= 'z') { - int e= si_prefixes[*next - 'E']; + int e= si_prefixes[*next - 'E'].exp; if (e) { if (next[1] == 'i') { - d*= pow( 2, e/0.3); + d*= si_prefixes[*next - 'E'].bin_val; next+=2; } else { - d*= pow(10, e); + d*= si_prefixes[*next - 'E'].dec_val; next++; } } @@ -129,10 +150,11 @@ struct AVExpr { enum { e_value, e_const, e_func0, e_func1, e_func2, e_squish, e_gauss, e_ld, e_isnan, e_isinf, - e_mod, e_max, e_min, e_eq, e_gt, e_gte, + e_mod, e_max, e_min, e_eq, e_gt, e_gte, e_lte, e_lt, e_pow, e_mul, e_div, e_add, - e_last, e_st, e_while, e_floor, e_ceil, e_trunc, - e_sqrt, e_not, + e_last, e_st, e_while, e_taylor, e_root, e_floor, e_ceil, e_trunc, + e_sqrt, e_not, e_random, e_hypot, e_gcd, + e_if, e_ifnot, e_print, e_bitand, e_bitor, e_between, e_clip } type; double value; // is sign in other types union { @@ -141,9 +163,15 @@ struct AVExpr { double (*func1)(void *, double); double (*func2)(void *, double, double); } a; - struct AVExpr *param[2]; + struct AVExpr *param[3]; + double *var; }; +static double etime(double v) +{ + return av_gettime() * 0.000001; +} + static double eval_expr(Parser *p, AVExpr *e) { switch (e->type) { @@ -161,29 +189,124 @@ static double eval_expr(Parser *p, AVExpr *e) case e_ceil : return e->value * ceil (eval_expr(p, e->param[0])); case e_trunc: return e->value * trunc(eval_expr(p, e->param[0])); case e_sqrt: return e->value * sqrt (eval_expr(p, e->param[0])); - case e_not: return e->value * eval_expr(p, e->param[0]) == 0; + case e_not: return e->value * (eval_expr(p, e->param[0]) == 0); + case e_if: return e->value * (eval_expr(p, e->param[0]) ? eval_expr(p, e->param[1]) : + e->param[2] ? eval_expr(p, e->param[2]) : 0); + case e_ifnot: return e->value * (!eval_expr(p, e->param[0]) ? eval_expr(p, e->param[1]) : + e->param[2] ? eval_expr(p, e->param[2]) : 0); + case e_clip: { + double x = eval_expr(p, e->param[0]); + double min = eval_expr(p, e->param[1]), max = eval_expr(p, e->param[2]); + if (isnan(min) || isnan(max) || isnan(x) || min > max) + return NAN; + return e->value * av_clipd(eval_expr(p, e->param[0]), min, max); + } + case e_between: { + double d = eval_expr(p, e->param[0]); + return e->value * (d >= eval_expr(p, e->param[1]) && + d <= eval_expr(p, e->param[2])); + } + case e_print: { + double x = eval_expr(p, e->param[0]); + int level = e->param[1] ? av_clip(eval_expr(p, e->param[1]), INT_MIN, INT_MAX) : AV_LOG_INFO; + av_log(p, level, "%f\n", x); + return x; + } + case e_random:{ + int idx= av_clip(eval_expr(p, e->param[0]), 0, VARS-1); + uint64_t r= isnan(p->var[idx]) ? 0 : p->var[idx]; + r= r*1664525+1013904223; + p->var[idx]= r; + return e->value * (r * (1.0/UINT64_MAX)); + } case e_while: { double d = NAN; while (eval_expr(p, e->param[0])) d=eval_expr(p, e->param[1]); return d; } + case e_taylor: { + double t = 1, d = 0, v; + double x = eval_expr(p, e->param[1]); + int id = e->param[2] ? av_clip(eval_expr(p, e->param[2]), 0, VARS-1) : 0; + int i; + double var0 = p->var[id]; + for(i=0; i<1000; i++) { + double ld = d; + p->var[id] = i; + v = eval_expr(p, e->param[0]); + d += t*v; + if(ld==d && v) + break; + t *= x / (i+1); + } + p->var[id] = var0; + return d; + } + case e_root: { + int i, j; + double low = -1, high = -1, v, low_v = -DBL_MAX, high_v = DBL_MAX; + double var0 = p->var[0]; + double x_max = eval_expr(p, e->param[1]); + for(i=-1; i<1024; i++) { + if(i<255) { + p->var[0] = ff_reverse[i&255]*x_max/255; + } else { + p->var[0] = x_max*pow(0.9, i-255); + if (i&1) p->var[0] *= -1; + if (i&2) p->var[0] += low; + else p->var[0] += high; + } + v = eval_expr(p, e->param[0]); + if (v<=0 && v>low_v) { + low = p->var[0]; + low_v = v; + } + if (v>=0 && v<high_v) { + high = p->var[0]; + high_v = v; + } + if (low>=0 && high>=0){ + for (j=0; j<1000; j++) { + p->var[0] = (low+high)*0.5; + if (low == p->var[0] || high == p->var[0]) + break; + v = eval_expr(p, e->param[0]); + if (v<=0) low = p->var[0]; + if (v>=0) high= p->var[0]; + if (isnan(v)) { + low = high = v; + break; + } + } + break; + } + } + p->var[0] = var0; + return -low_v<high_v ? low : high; + } default: { double d = eval_expr(p, e->param[0]); double d2 = eval_expr(p, e->param[1]); switch (e->type) { - case e_mod: return e->value * (d - floor(d/d2)*d2); + case e_mod: return e->value * (d - floor((!CONFIG_FTRAPV || d2) ? d / d2 : d * INFINITY) * d2); + case e_gcd: return e->value * av_gcd(d,d2); case e_max: return e->value * (d > d2 ? d : d2); case e_min: return e->value * (d < d2 ? d : d2); case e_eq: return e->value * (d == d2 ? 1.0 : 0.0); case e_gt: return e->value * (d > d2 ? 1.0 : 0.0); case e_gte: return e->value * (d >= d2 ? 1.0 : 0.0); + case e_lt: return e->value * (d < d2 ? 1.0 : 0.0); + case e_lte: return e->value * (d <= d2 ? 1.0 : 0.0); case e_pow: return e->value * pow(d, d2); case e_mul: return e->value * (d * d2); - case e_div: return e->value * (d / d2); + case e_div: return e->value * ((!CONFIG_FTRAPV || d2 ) ? (d / d2) : d * INFINITY); case e_add: return e->value * (d + d2); case e_last:return e->value * d2; case e_st : return e->value * (p->var[av_clip(d, 0, VARS-1)]= d2); + case e_hypot:return e->value * hypot(d, d2); + case e_bitand: return isnan(d) || isnan(d2) ? NAN : e->value * ((long int)d & (long int)d2); + case e_bitor: return isnan(d) || isnan(d2) ? NAN : e->value * ((long int)d | (long int)d2); } } } @@ -197,6 +320,8 @@ void av_expr_free(AVExpr *e) if (!e) return; av_expr_free(e->param[0]); av_expr_free(e->param[1]); + av_expr_free(e->param[2]); + av_freep(&e->var); av_freep(&e); } @@ -229,6 +354,15 @@ static int parse_primary(AVExpr **e, Parser *p) return 0; } } + for (i = 0; i < FF_ARRAY_ELEMS(constants); i++) { + if (strmatch(p->s, constants[i].name)) { + p->s += strlen(constants[i].name); + d->type = e_value; + d->value = constants[i].value; + *e = d; + return 0; + } + } p->s= strchr(p->s, '('); if (!p->s) { @@ -259,6 +393,10 @@ static int parse_primary(AVExpr **e, Parser *p) p->s++; // "," parse_expr(&d->param[1], p); } + if (p->s[0]== ',') { + p->s++; // "," + parse_expr(&d->param[2], p); + } if (p->s[0] != ')') { av_log(p, AV_LOG_ERROR, "Missing ')' or too many args in '%s'\n", s0); av_expr_free(d); @@ -279,6 +417,7 @@ static int parse_primary(AVExpr **e, Parser *p) else if (strmatch(next, "exp" )) d->a.func0 = exp; else if (strmatch(next, "log" )) d->a.func0 = log; else if (strmatch(next, "abs" )) d->a.func0 = fabs; + else if (strmatch(next, "time" )) d->a.func0 = etime; else if (strmatch(next, "squish")) d->type = e_squish; else if (strmatch(next, "gauss" )) d->type = e_gauss; else if (strmatch(next, "mod" )) d->type = e_mod; @@ -287,18 +426,31 @@ static int parse_primary(AVExpr **e, Parser *p) else if (strmatch(next, "eq" )) d->type = e_eq; else if (strmatch(next, "gte" )) d->type = e_gte; else if (strmatch(next, "gt" )) d->type = e_gt; - else if (strmatch(next, "lte" )) { AVExpr *tmp = d->param[1]; d->param[1] = d->param[0]; d->param[0] = tmp; d->type = e_gte; } - else if (strmatch(next, "lt" )) { AVExpr *tmp = d->param[1]; d->param[1] = d->param[0]; d->param[0] = tmp; d->type = e_gt; } + else if (strmatch(next, "lte" )) d->type = e_lte; + else if (strmatch(next, "lt" )) d->type = e_lt; else if (strmatch(next, "ld" )) d->type = e_ld; else if (strmatch(next, "isnan" )) d->type = e_isnan; else if (strmatch(next, "isinf" )) d->type = e_isinf; else if (strmatch(next, "st" )) d->type = e_st; else if (strmatch(next, "while" )) d->type = e_while; + else if (strmatch(next, "taylor")) d->type = e_taylor; + else if (strmatch(next, "root" )) d->type = e_root; else if (strmatch(next, "floor" )) d->type = e_floor; else if (strmatch(next, "ceil" )) d->type = e_ceil; else if (strmatch(next, "trunc" )) d->type = e_trunc; else if (strmatch(next, "sqrt" )) d->type = e_sqrt; else if (strmatch(next, "not" )) d->type = e_not; + else if (strmatch(next, "pow" )) d->type = e_pow; + else if (strmatch(next, "print" )) d->type = e_print; + else if (strmatch(next, "random")) d->type = e_random; + else if (strmatch(next, "hypot" )) d->type = e_hypot; + else if (strmatch(next, "gcd" )) d->type = e_gcd; + else if (strmatch(next, "if" )) d->type = e_if; + else if (strmatch(next, "ifnot" )) d->type = e_ifnot; + else if (strmatch(next, "bitand")) d->type = e_bitand; + else if (strmatch(next, "bitor" )) d->type = e_bitor; + else if (strmatch(next, "between"))d->type = e_between; + else if (strmatch(next, "clip" )) d->type = e_clip; else { for (i=0; p->func1_names && p->func1_names[i]; i++) { if (strmatch(next, p->func1_names[i])) { @@ -327,7 +479,7 @@ static int parse_primary(AVExpr **e, Parser *p) return 0; } -static AVExpr *new_eval_expr(int type, int value, AVExpr *p0, AVExpr *p1) +static AVExpr *make_eval_expr(int type, int value, AVExpr *p0, AVExpr *p1) { AVExpr *e = av_mallocz(sizeof(AVExpr)); if (!e) @@ -374,7 +526,7 @@ static int parse_factor(AVExpr **e, Parser *p) av_expr_free(e1); return ret; } - e0 = new_eval_expr(e_pow, 1, e1, e2); + e0 = make_eval_expr(e_pow, 1, e1, e2); if (!e0) { av_expr_free(e1); av_expr_free(e2); @@ -401,7 +553,7 @@ static int parse_term(AVExpr **e, Parser *p) av_expr_free(e1); return ret; } - e0 = new_eval_expr(c == '*' ? e_mul : e_div, 1, e1, e2); + e0 = make_eval_expr(c == '*' ? e_mul : e_div, 1, e1, e2); if (!e0) { av_expr_free(e1); av_expr_free(e2); @@ -424,7 +576,7 @@ static int parse_subexpr(AVExpr **e, Parser *p) av_expr_free(e1); return ret; } - e0 = new_eval_expr(e_add, 1, e1, e2); + e0 = make_eval_expr(e_add, 1, e1, e2); if (!e0) { av_expr_free(e1); av_expr_free(e2); @@ -453,7 +605,7 @@ static int parse_expr(AVExpr **e, Parser *p) av_expr_free(e1); return ret; } - e0 = new_eval_expr(e_last, 1, e1, e2); + e0 = make_eval_expr(e_last, 1, e1, e2); if (!e0) { av_expr_free(e1); av_expr_free(e2); @@ -484,8 +636,22 @@ static int verify_expr(AVExpr *e) case e_trunc: case e_sqrt: case e_not: - return verify_expr(e->param[0]); - default: return verify_expr(e->param[0]) && verify_expr(e->param[1]); + case e_random: + return verify_expr(e->param[0]) && !e->param[1]; + case e_print: + return verify_expr(e->param[0]) + && (!e->param[1] || verify_expr(e->param[1])); + case e_if: + case e_ifnot: + case e_taylor: + return verify_expr(e->param[0]) && verify_expr(e->param[1]) + && (!e->param[2] || verify_expr(e->param[2])); + case e_between: + case e_clip: + return verify_expr(e->param[0]) && + verify_expr(e->param[1]) && + verify_expr(e->param[2]); + default: return verify_expr(e->param[0]) && verify_expr(e->param[1]) && !e->param[2]; } } @@ -509,7 +675,7 @@ int av_expr_parse(AVExpr **expr, const char *s, if (!av_isspace(*s++)) *wp++ = s[-1]; *wp++ = 0; - p.class = &class; + p.class = &eval_class; p.stack_index=100; p.s= w; p.const_names = const_names; @@ -523,18 +689,23 @@ int av_expr_parse(AVExpr **expr, const char *s, if ((ret = parse_expr(&e, &p)) < 0) goto end; if (*p.s) { - av_expr_free(e); av_log(&p, AV_LOG_ERROR, "Invalid chars '%s' at the end of expression '%s'\n", p.s, s0); ret = AVERROR(EINVAL); goto end; } if (!verify_expr(e)) { - av_expr_free(e); ret = AVERROR(EINVAL); goto end; } + e->var= av_mallocz(sizeof(double) *VARS); + if (!e->var) { + ret = AVERROR(ENOMEM); + goto end; + } *expr = e; + e = NULL; end: + av_expr_free(e); av_free(w); return ret; } @@ -542,6 +713,7 @@ end: double av_expr_eval(AVExpr *e, const double *const_values, void *opaque) { Parser p = { 0 }; + p.var= e->var; p.const_values = const_values; p.opaque = opaque; diff --git a/libavutil/eval.h b/libavutil/eval.h index ccb29e7a33..dacd22b96e 100644 --- a/libavutil/eval.h +++ b/libavutil/eval.h @@ -1,20 +1,20 @@ /* * Copyright (c) 2002 Michael Niedermayer <michaelni@gmx.at> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -45,7 +45,7 @@ typedef struct AVExpr AVExpr; * @param funcs2 NULL terminated array of function pointers for functions which take 2 arguments * @param opaque a pointer which will be passed to all functions from funcs1 and funcs2 * @param log_ctx parent logging context - * @return 0 in case of success, a negative value corresponding to an + * @return >= 0 in case of success, a negative value corresponding to an * AVERROR code otherwise */ int av_expr_parse_and_eval(double *res, const char *s, @@ -68,7 +68,7 @@ int av_expr_parse_and_eval(double *res, const char *s, * @param func2_names NULL terminated array of zero terminated strings of funcs2 identifiers * @param funcs2 NULL terminated array of function pointers for functions which take 2 arguments * @param log_ctx parent logging context - * @return 0 in case of success, a negative value corresponding to an + * @return >= 0 in case of success, a negative value corresponding to an * AVERROR code otherwise */ int av_expr_parse(AVExpr **expr, const char *s, @@ -102,7 +102,7 @@ void av_expr_free(AVExpr *e); * @param numstr a string representing a number, may contain one of * the International System number postfixes, for example 'K', 'M', * 'G'. If 'i' is appended after the postfix, powers of 2 are used - * instead of powers of 10. The 'B' postfix multiplies the value for + * instead of powers of 10. The 'B' postfix multiplies the value by * 8, and can be appended after another postfix or used alone. This * allows using for example 'KB', 'MiB', 'G' and 'B' as postfix. * @param tail if non-NULL puts here the pointer to the char next diff --git a/libavutil/ffmath.h b/libavutil/ffmath.h new file mode 100644 index 0000000000..aad1347f26 --- /dev/null +++ b/libavutil/ffmath.h @@ -0,0 +1,67 @@ +/* + * copyright (c) 2016 Ganesh Ajjanagadde <gajjanag@gmail.com> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +/** + * @file + * internal math functions header + */ + +#ifndef AVUTIL_FFMATH_H +#define AVUTIL_FFMATH_H + +#include "attributes.h" +#include "libm.h" + +/** + * Compute 10^x for floating point values. Note: this function is by no means + * "correctly rounded", and is meant as a fast, reasonably accurate approximation. + * For instance, maximum relative error for the double precision variant is + * ~ 1e-13 for very small and very large values. + * This is ~2x faster than GNU libm's approach, which is still off by 2ulp on + * some inputs. + * @param x exponent + * @return 10^x + */ +static av_always_inline double ff_exp10(double x) +{ + return exp2(M_LOG2_10 * x); +} + +static av_always_inline float ff_exp10f(float x) +{ + return exp2f(M_LOG2_10 * x); +} + +/** + * Compute x^y for floating point x, y. Note: this function is faster than the + * libm variant due to mainly 2 reasons: + * 1. It does not handle any edge cases. In particular, this is only guaranteed + * to work correctly for x > 0. + * 2. It is not as accurate as a standard nearly "correctly rounded" libm variant. + * @param x base + * @param y exponent + * @return x^y + */ +static av_always_inline float ff_fast_powf(float x, float y) +{ + return expf(logf(x) * y); +} + +#endif /* AVUTIL_FFMATH_H */ diff --git a/libavutil/fifo-test.c b/libavutil/fifo-test.c index f6615b1034..71a9172a45 100644 --- a/libavutil/fifo-test.c +++ b/libavutil/fifo-test.c @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -38,6 +38,14 @@ int main(void) } printf("\n"); + /* peek_at at FIFO */ + n = av_fifo_size(fifo) / sizeof(int); + for (i = 0; i < n; i++) { + av_fifo_generic_peek_at(fifo, &j, i * sizeof(int), sizeof(j), NULL); + printf("%d: %d\n", i, j); + } + printf("\n"); + /* read data */ for (i = 0; av_fifo_size(fifo) >= sizeof(int); i++) { av_fifo_generic_read(fifo, &j, sizeof(int), NULL); @@ -45,6 +53,21 @@ int main(void) } printf("\n"); + /* test *ndx overflow */ + av_fifo_reset(fifo); + fifo->rndx = fifo->wndx = ~(uint32_t)0 - 5; + + /* fill data */ + for (i = 0; av_fifo_space(fifo) >= sizeof(int); i++) + av_fifo_generic_write(fifo, &i, sizeof(int), NULL); + + /* peek_at at FIFO */ + n = av_fifo_size(fifo) / sizeof(int); + for (i = 0; i < n; i++) { + av_fifo_generic_peek_at(fifo, &j, i * sizeof(int), sizeof(j), NULL); + printf("%d: %d\n", i, j); + } + av_fifo_free(fifo); return 0; diff --git a/libavutil/fifo.c b/libavutil/fifo.c index a42899c2b3..986729aedb 100644 --- a/libavutil/fifo.c +++ b/libavutil/fifo.c @@ -3,58 +3,83 @@ * Copyright (c) 2000, 2001, 2002 Fabrice Bellard * Copyright (c) 2006 Roman Shaposhnik * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ + +#include "avassert.h" #include "common.h" #include "fifo.h" -AVFifoBuffer *av_fifo_alloc(unsigned int size) +static AVFifoBuffer *fifo_alloc_common(void *buffer, size_t size) { - AVFifoBuffer *f = av_mallocz(sizeof(AVFifoBuffer)); - if (!f) + AVFifoBuffer *f; + if (!buffer) return NULL; - f->buffer = av_malloc(size); + f = av_mallocz(sizeof(AVFifoBuffer)); + if (!f) { + av_free(buffer); + return NULL; + } + f->buffer = buffer; f->end = f->buffer + size; av_fifo_reset(f); - if (!f->buffer) - av_freep(&f); return f; } +AVFifoBuffer *av_fifo_alloc(unsigned int size) +{ + void *buffer = av_malloc(size); + return fifo_alloc_common(buffer, size); +} + +AVFifoBuffer *av_fifo_alloc_array(size_t nmemb, size_t size) +{ + void *buffer = av_malloc_array(nmemb, size); + return fifo_alloc_common(buffer, nmemb * size); +} + void av_fifo_free(AVFifoBuffer *f) { if (f) { - av_free(f->buffer); + av_freep(&f->buffer); av_free(f); } } +void av_fifo_freep(AVFifoBuffer **f) +{ + if (f) { + av_fifo_free(*f); + *f = NULL; + } +} + void av_fifo_reset(AVFifoBuffer *f) { f->wptr = f->rptr = f->buffer; f->wndx = f->rndx = 0; } -int av_fifo_size(AVFifoBuffer *f) +int av_fifo_size(const AVFifoBuffer *f) { return (uint32_t)(f->wndx - f->rndx); } -int av_fifo_space(AVFifoBuffer *f) +int av_fifo_space(const AVFifoBuffer *f) { return f->end - f->buffer - av_fifo_size(f); } @@ -68,7 +93,7 @@ int av_fifo_realloc2(AVFifoBuffer *f, unsigned int new_size) AVFifoBuffer *f2 = av_fifo_alloc(new_size); if (!f2) - return -1; + return AVERROR(ENOMEM); av_fifo_generic_read(f, f2->buffer, len, NULL); f2->wptr += len; f2->wndx += len; @@ -79,31 +104,112 @@ int av_fifo_realloc2(AVFifoBuffer *f, unsigned int new_size) return 0; } +int av_fifo_grow(AVFifoBuffer *f, unsigned int size) +{ + unsigned int old_size = f->end - f->buffer; + if(size + (unsigned)av_fifo_size(f) < size) + return AVERROR(EINVAL); + + size += av_fifo_size(f); + + if (old_size < size) + return av_fifo_realloc2(f, FFMAX(size, 2*size)); + return 0; +} + /* src must NOT be const as it can be a context for func that may need * updating (like a pointer or byte counter) */ int av_fifo_generic_write(AVFifoBuffer *f, void *src, int size, int (*func)(void *, void *, int)) { int total = size; + uint32_t wndx= f->wndx; + uint8_t *wptr= f->wptr; + do { - int len = FFMIN(f->end - f->wptr, size); + int len = FFMIN(f->end - wptr, size); if (func) { - if (func(src, f->wptr, len) <= 0) + len = func(src, wptr, len); + if (len <= 0) break; } else { - memcpy(f->wptr, src, len); + memcpy(wptr, src, len); src = (uint8_t *)src + len; } // Write memory barrier needed for SMP here in theory - f->wptr += len; - if (f->wptr >= f->end) - f->wptr = f->buffer; - f->wndx += len; + wptr += len; + if (wptr >= f->end) + wptr = f->buffer; + wndx += len; size -= len; } while (size > 0); + f->wndx= wndx; + f->wptr= wptr; return total - size; } +int av_fifo_generic_peek_at(AVFifoBuffer *f, void *dest, int offset, int buf_size, void (*func)(void*, void*, int)) +{ + uint8_t *rptr = f->rptr; + + av_assert2(offset >= 0); + + /* + * *ndx are indexes modulo 2^32, they are intended to overflow, + * to handle *ndx greater than 4gb. + */ + av_assert2(buf_size + (unsigned)offset <= f->wndx - f->rndx); + + if (offset >= f->end - rptr) + rptr += offset - (f->end - f->buffer); + else + rptr += offset; + + while (buf_size > 0) { + int len; + + if (rptr >= f->end) + rptr -= f->end - f->buffer; + + len = FFMIN(f->end - rptr, buf_size); + if (func) + func(dest, rptr, len); + else { + memcpy(dest, rptr, len); + dest = (uint8_t *)dest + len; + } + + buf_size -= len; + rptr += len; + } + + return 0; +} + +int av_fifo_generic_peek(AVFifoBuffer *f, void *dest, int buf_size, + void (*func)(void *, void *, int)) +{ +// Read memory barrier needed for SMP here in theory + uint8_t *rptr = f->rptr; + + do { + int len = FFMIN(f->end - rptr, buf_size); + if (func) + func(dest, rptr, len); + else { + memcpy(dest, rptr, len); + dest = (uint8_t *)dest + len; + } +// memory barrier needed for SMP here in theory + rptr += len; + if (rptr >= f->end) + rptr -= f->end - f->buffer; + buf_size -= len; + } while (buf_size > 0); + + return 0; +} + int av_fifo_generic_read(AVFifoBuffer *f, void *dest, int buf_size, void (*func)(void *, void *, int)) { @@ -126,6 +232,7 @@ int av_fifo_generic_read(AVFifoBuffer *f, void *dest, int buf_size, /** Discard data from the FIFO. */ void av_fifo_drain(AVFifoBuffer *f, int size) { + av_assert2(av_fifo_size(f) >= size); f->rptr += size; if (f->rptr >= f->end) f->rptr -= f->end - f->buffer; diff --git a/libavutil/fifo.h b/libavutil/fifo.h index ea30f5d2bd..dc7bc6f0dd 100644 --- a/libavutil/fifo.h +++ b/libavutil/fifo.h @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -42,12 +42,26 @@ typedef struct AVFifoBuffer { AVFifoBuffer *av_fifo_alloc(unsigned int size); /** + * Initialize an AVFifoBuffer. + * @param nmemb number of elements + * @param size size of the single element + * @return AVFifoBuffer or NULL in case of memory allocation failure + */ +AVFifoBuffer *av_fifo_alloc_array(size_t nmemb, size_t size); + +/** * Free an AVFifoBuffer. * @param f AVFifoBuffer to free */ void av_fifo_free(AVFifoBuffer *f); /** + * Free an AVFifoBuffer and reset pointer to NULL. + * @param f AVFifoBuffer to free + */ +void av_fifo_freep(AVFifoBuffer **f); + +/** * Reset the AVFifoBuffer to the state right after av_fifo_alloc, in particular it is emptied. * @param f AVFifoBuffer to reset */ @@ -59,7 +73,7 @@ void av_fifo_reset(AVFifoBuffer *f); * @param f AVFifoBuffer to read from * @return size */ -int av_fifo_size(AVFifoBuffer *f); +int av_fifo_size(const AVFifoBuffer *f); /** * Return the amount of space in bytes in the AVFifoBuffer, that is the @@ -67,7 +81,28 @@ int av_fifo_size(AVFifoBuffer *f); * @param f AVFifoBuffer to write into * @return size */ -int av_fifo_space(AVFifoBuffer *f); +int av_fifo_space(const AVFifoBuffer *f); + +/** + * Feed data at specific position from an AVFifoBuffer to a user-supplied callback. + * Similar as av_fifo_gereric_read but without discarding data. + * @param f AVFifoBuffer to read from + * @param offset offset from current read position + * @param buf_size number of bytes to read + * @param func generic read function + * @param dest data destination + */ +int av_fifo_generic_peek_at(AVFifoBuffer *f, void *dest, int offset, int buf_size, void (*func)(void*, void*, int)); + +/** + * Feed data from an AVFifoBuffer to a user-supplied callback. + * Similar as av_fifo_gereric_read but without discarding data. + * @param f AVFifoBuffer to read from + * @param buf_size number of bytes to read + * @param func generic read function + * @param dest data destination + */ +int av_fifo_generic_peek(AVFifoBuffer *f, void *dest, int buf_size, void (*func)(void*, void*, int)); /** * Feed data from an AVFifoBuffer to a user-supplied callback. @@ -95,6 +130,8 @@ int av_fifo_generic_write(AVFifoBuffer *f, void *src, int size, int (*func)(void /** * Resize an AVFifoBuffer. + * In case of reallocation failure, the old FIFO is kept unchanged. + * * @param f AVFifoBuffer to resize * @param size new AVFifoBuffer size in bytes * @return <0 for failure, >=0 otherwise @@ -102,6 +139,17 @@ int av_fifo_generic_write(AVFifoBuffer *f, void *src, int size, int (*func)(void int av_fifo_realloc2(AVFifoBuffer *f, unsigned int size); /** + * Enlarge an AVFifoBuffer. + * In case of reallocation failure, the old FIFO is kept unchanged. + * The new fifo size may be larger than the requested size. + * + * @param f AVFifoBuffer to resize + * @param additional_space the amount of space in bytes to allocate in addition to av_fifo_size() + * @return <0 for failure, >=0 otherwise + */ +int av_fifo_grow(AVFifoBuffer *f, unsigned int additional_space); + +/** * Read and discard the specified amount of data from an AVFifoBuffer. * @param f AVFifoBuffer to read from * @param size amount of data to read in bytes diff --git a/libavutil/file-test.c b/libavutil/file-test.c new file mode 100644 index 0000000000..0324081fb9 --- /dev/null +++ b/libavutil/file-test.c @@ -0,0 +1,34 @@ +/* + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "file.c" + +#undef printf + +int main(void) +{ + uint8_t *buf; + size_t size; + if (av_file_map("file.c", &buf, &size, 0, NULL) < 0) + return 1; + + buf[0] = 's'; + printf("%s", buf); + av_file_unmap(buf, size); + return 0; +} diff --git a/libavutil/file.c b/libavutil/file.c index 18a110aeeb..7bdf6cde84 100644 --- a/libavutil/file.c +++ b/libavutil/file.c @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -136,3 +136,7 @@ void av_file_unmap(uint8_t *bufptr, size_t size) av_free(bufptr); #endif } + +int av_tempfile(const char *prefix, char **filename, int log_offset, void *log_ctx) { + return avpriv_tempfile(prefix, filename, log_offset, log_ctx); +} diff --git a/libavutil/file.h b/libavutil/file.h index e3f02a8308..8666c7b1d5 100644 --- a/libavutil/file.h +++ b/libavutil/file.h @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -40,6 +40,7 @@ * @return a non negative number in case of success, a negative value * corresponding to an AVERROR error code in case of failure */ +av_warn_unused_result int av_file_map(const char *filename, uint8_t **bufptr, size_t *size, int log_offset, void *log_ctx); @@ -51,4 +52,18 @@ int av_file_map(const char *filename, uint8_t **bufptr, size_t *size, */ void av_file_unmap(uint8_t *bufptr, size_t size); +/** + * Wrapper to work around the lack of mkstemp() on mingw. + * Also, tries to create file in /tmp first, if possible. + * *prefix can be a character constant; *filename will be allocated internally. + * @return file descriptor of opened file (or negative value corresponding to an + * AVERROR code on error) + * and opened file name in **filename. + * @note On very old libcs it is necessary to set a secure umask before + * calling this, av_tempfile() can't call umask itself as it is used in + * libraries and could interfere with the calling application. + * @deprecated as fd numbers cannot be passed saftely between libs on some platforms + */ +int av_tempfile(const char *prefix, char **filename, int log_offset, void *log_ctx); + #endif /* AVUTIL_FILE_H */ diff --git a/libavutil/file_open.c b/libavutil/file_open.c index f14ea70862..6e58cc16e6 100644 --- a/libavutil/file_open.c +++ b/libavutil/file_open.c @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -57,7 +57,7 @@ static int win32_open(const char *filename_utf8, int oflag, int pmode) return fd; fallback: - /* filename may be be in CP_ACP */ + /* filename may be in CP_ACP */ return _sopen(filename_utf8, oflag, SH_DENYNO, pmode); } #define open win32_open @@ -77,12 +77,110 @@ int avpriv_open(const char *filename, int flags, ...) #ifdef O_CLOEXEC flags |= O_CLOEXEC; #endif +#ifdef O_NOINHERIT + flags |= O_NOINHERIT; +#endif fd = open(filename, flags, mode); #if HAVE_FCNTL - if (fd != -1) - fcntl(fd, F_SETFD, FD_CLOEXEC); + if (fd != -1) { + if (fcntl(fd, F_SETFD, FD_CLOEXEC) == -1) + av_log(NULL, AV_LOG_DEBUG, "Failed to set close on exec\n"); + } #endif return fd; } + +typedef struct FileLogContext { + const AVClass *class; + int log_offset; + void *log_ctx; +} FileLogContext; + +static const AVClass file_log_ctx_class = { + "TEMPFILE", av_default_item_name, NULL, LIBAVUTIL_VERSION_INT, + offsetof(FileLogContext, log_offset), offsetof(FileLogContext, log_ctx) +}; + +int avpriv_tempfile(const char *prefix, char **filename, int log_offset, void *log_ctx) +{ + FileLogContext file_log_ctx = { &file_log_ctx_class, log_offset, log_ctx }; + int fd = -1; +#if !HAVE_MKSTEMP + void *ptr= tempnam(NULL, prefix); + if(!ptr) + ptr= tempnam(".", prefix); + *filename = av_strdup(ptr); +#undef free + free(ptr); +#else + size_t len = strlen(prefix) + 12; /* room for "/tmp/" and "XXXXXX\0" */ + *filename = av_malloc(len); +#endif + /* -----common section-----*/ + if (!*filename) { + av_log(&file_log_ctx, AV_LOG_ERROR, "ff_tempfile: Cannot allocate file name\n"); + return AVERROR(ENOMEM); + } +#if !HAVE_MKSTEMP +# ifndef O_BINARY +# define O_BINARY 0 +# endif +# ifndef O_EXCL +# define O_EXCL 0 +# endif + fd = open(*filename, O_RDWR | O_BINARY | O_CREAT | O_EXCL, 0600); +#else + snprintf(*filename, len, "/tmp/%sXXXXXX", prefix); + fd = mkstemp(*filename); +#ifdef _WIN32 + if (fd < 0) { + snprintf(*filename, len, "./%sXXXXXX", prefix); + fd = mkstemp(*filename); + } +#endif +#endif + /* -----common section-----*/ + if (fd < 0) { + int err = AVERROR(errno); + av_log(&file_log_ctx, AV_LOG_ERROR, "ff_tempfile: Cannot open temporary file %s\n", *filename); + av_freep(filename); + return err; + } + return fd; /* success */ +} + +FILE *av_fopen_utf8(const char *path, const char *mode) +{ + int fd; + int access; + const char *m = mode; + + switch (*m++) { + case 'r': access = O_RDONLY; break; + case 'w': access = O_CREAT|O_WRONLY|O_TRUNC; break; + case 'a': access = O_CREAT|O_WRONLY|O_APPEND; break; + default : + errno = EINVAL; + return NULL; + } + while (*m) { + if (*m == '+') { + access &= ~(O_RDONLY | O_WRONLY); + access |= O_RDWR; + } else if (*m == 'b') { +#ifdef O_BINARY + access |= O_BINARY; +#endif + } else if (*m) { + errno = EINVAL; + return NULL; + } + m++; + } + fd = avpriv_open(path, access, 0666); + if (fd == -1) + return NULL; + return fdopen(fd, mode); +} diff --git a/libavutil/fixed_dsp.c b/libavutil/fixed_dsp.c new file mode 100644 index 0000000000..8c018581df --- /dev/null +++ b/libavutil/fixed_dsp.c @@ -0,0 +1,167 @@ +/* + * Copyright (c) 2012 + * MIPS Technologies, Inc., California. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * 3. Neither the name of the MIPS Technologies, Inc., nor the names of its + * contributors may be used to endorse or promote products derived from + * this software without specific prior written permission. + * + * THIS SOFTWARE IS PROVIDED BY THE MIPS TECHNOLOGIES, INC. ``AS IS'' AND + * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE + * ARE DISCLAIMED. IN NO EVENT SHALL THE MIPS TECHNOLOGIES, INC. BE LIABLE + * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT + * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY + * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF + * SUCH DAMAGE. + * + * Author: Nedeljko Babic (nedeljko.babic imgtec com) + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "fixed_dsp.h" + +static void vector_fmul_add_c(int *dst, const int *src0, const int *src1, const int *src2, int len){ + int i; + int64_t accu; + + for (i=0; i<len; i++) { + accu = (int64_t)src0[i] * src1[i]; + dst[i] = src2[i] + (int)((accu + 0x40000000) >> 31); + } +} + +static void vector_fmul_reverse_c(int *dst, const int *src0, const int *src1, int len) +{ + int i; + int64_t accu; + + src1 += len-1; + for (i=0; i<len; i++) { + accu = (int64_t)src0[i] * src1[-i]; + dst[i] = (int)((accu+0x40000000) >> 31); + } +} + +static void vector_fmul_window_scaled_c(int16_t *dst, const int32_t *src0, + const int32_t *src1, const int32_t *win, + int len, uint8_t bits) +{ + int32_t s0, s1, wi, wj, i,j, round; + + dst += len; + win += len; + src0+= len; + round = bits? 1 << (bits-1) : 0; + + for (i=-len, j=len-1; i<0; i++, j--) { + s0 = src0[i]; + s1 = src1[j]; + wi = win[i]; + wj = win[j]; + dst[i] = av_clip_int16(((((int64_t)s0*wj - (int64_t)s1*wi + 0x40000000) >> 31) + round) >> bits); + dst[j] = av_clip_int16(((((int64_t)s0*wi + (int64_t)s1*wj + 0x40000000) >> 31) + round) >> bits); + } +} + +static void vector_fmul_window_c(int32_t *dst, const int32_t *src0, + const int32_t *src1, const int32_t *win, + int len) +{ + int32_t s0, s1, wi, wj, i, j; + + dst += len; + win += len; + src0+= len; + + for (i=-len, j=len-1; i<0; i++, j--) { + s0 = src0[i]; + s1 = src1[j]; + wi = win[i]; + wj = win[j]; + dst[i] = ((int64_t)s0*wj - (int64_t)s1*wi + 0x40000000) >> 31; + dst[j] = ((int64_t)s0*wi + (int64_t)s1*wj + 0x40000000) >> 31; + } +} + +static void vector_fmul_c(int *dst, const int *src0, const int *src1, int len) +{ + int i; + int64_t accu; + + for (i = 0; i < len; i++){ + accu = (int64_t)src0[i] * src1[i]; + dst[i] = (int)((accu+0x40000000) >> 31); + } +} + +static int scalarproduct_fixed_c(const int *v1, const int *v2, int len) +{ + /** p is initialized with 0x40000000 so that the proper rounding will occur + * at the end */ + int64_t p = 0x40000000; + int i; + + for (i = 0; i < len; i++) + p += (int64_t)v1[i] * v2[i]; + + return (int)(p >> 31); +} + +static void butterflies_fixed_c(int *v1, int *v2, int len) +{ + int i; + + for (i = 0; i < len; i++){ + int t = v1[i] - v2[i]; + v1[i] += v2[i]; + v2[i] = t; + } +} + +AVFixedDSPContext * avpriv_alloc_fixed_dsp(int bit_exact) +{ + AVFixedDSPContext * fdsp = av_malloc(sizeof(AVFixedDSPContext)); + + if (!fdsp) + return NULL; + + fdsp->vector_fmul_window_scaled = vector_fmul_window_scaled_c; + fdsp->vector_fmul_window = vector_fmul_window_c; + fdsp->vector_fmul = vector_fmul_c; + fdsp->vector_fmul_add = vector_fmul_add_c; + fdsp->vector_fmul_reverse = vector_fmul_reverse_c; + fdsp->butterflies_fixed = butterflies_fixed_c; + fdsp->scalarproduct_fixed = scalarproduct_fixed_c; + + if (ARCH_X86) + ff_fixed_dsp_init_x86(fdsp); + + return fdsp; +} diff --git a/libavutil/fixed_dsp.h b/libavutil/fixed_dsp.h new file mode 100644 index 0000000000..f554cb5038 --- /dev/null +++ b/libavutil/fixed_dsp.h @@ -0,0 +1,205 @@ +/* + * Copyright (c) 2012 + * MIPS Technologies, Inc., California. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * 3. Neither the name of the MIPS Technologies, Inc., nor the names of its + * contributors may be used to endorse or promote products derived from + * this software without specific prior written permission. + * + * THIS SOFTWARE IS PROVIDED BY THE MIPS TECHNOLOGIES, INC. ``AS IS'' AND + * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE + * ARE DISCLAIMED. IN NO EVENT SHALL THE MIPS TECHNOLOGIES, INC. BE LIABLE + * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT + * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY + * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF + * SUCH DAMAGE. + * + * Author: Nedeljko Babic (nbabic@mips.com) + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#ifndef AVUTIL_FIXED_DSP_H +#define AVUTIL_FIXED_DSP_H + +#include <stdint.h> +#include "attributes.h" +#include "common.h" +#include "libavcodec/mathops.h" + +typedef struct AVFixedDSPContext { + /* Assume len is a multiple of 16, and arrays are 32-byte aligned */ + /* Results of multiplications are scaled down by 31 bit (and rounded) if not + * stated otherwise */ + + /** + * Overlap/add with window function. + * Result is scaled down by "bits" bits. + * Used primarily by MDCT-based audio codecs. + * Source and destination vectors must overlap exactly or not at all. + * + * @param dst result vector + * constraints: 16-byte aligned + * @param src0 first source vector + * constraints: 16-byte aligned + * @param src1 second source vector + * constraints: 16-byte aligned + * @param win half-window vector + * constraints: 16-byte aligned + * @param len length of vector + * constraints: multiple of 4 + * @param bits scaling parameter + * + */ + void (*vector_fmul_window_scaled)(int16_t *dst, const int32_t *src0, const int32_t *src1, const int32_t *win, int len, uint8_t bits); + + /** + * Overlap/add with window function. + * Used primarily by MDCT-based audio codecs. + * Source and destination vectors must overlap exactly or not at all. + * + * @param dst result vector + * constraints: 32-byte aligned + * @param src0 first source vector + * constraints: 16-byte aligned + * @param src1 second source vector + * constraints: 16-byte aligned + * @param win half-window vector + * constraints: 16-byte aligned + * @param len length of vector + * constraints: multiple of 4 + */ + void (*vector_fmul_window)(int32_t *dst, const int32_t *src0, const int32_t *src1, const int32_t *win, int len); + + /** + * Fixed-point multiplication that calculates the entry wise product of two + * vectors of integers and stores the result in a vector of integers. + * + * @param dst output vector + * constraints: 32-byte aligned + * @param src0 first input vector + * constraints: 32-byte aligned + * @param src1 second input vector + * constraints: 32-byte aligned + * @param len number of elements in the input + * constraints: multiple of 16 + */ + void (*vector_fmul)(int *dst, const int *src0, const int *src1, + int len); + + void (*vector_fmul_reverse)(int *dst, const int *src0, const int *src1, int len); + /** + * Calculate the entry wise product of two vectors of integers, add a third vector of + * integers and store the result in a vector of integers. + * + * @param dst output vector + * constraints: 32-byte aligned + * @param src0 first input vector + * constraints: 32-byte aligned + * @param src1 second input vector + * constraints: 32-byte aligned + * @param src2 third input vector + * constraints: 32-byte aligned + * @param len number of elements in the input + * constraints: multiple of 16 + */ + void (*vector_fmul_add)(int *dst, const int *src0, const int *src1, + const int *src2, int len); + + /** + * Calculate the scalar product of two vectors of integers. + * + * @param v1 first vector, 16-byte aligned + * @param v2 second vector, 16-byte aligned + * @param len length of vectors, multiple of 4 + * + * @return sum of elementwise products + */ + int (*scalarproduct_fixed)(const int *v1, const int *v2, int len); + + /** + * Calculate the sum and difference of two vectors of integers. + * + * @param v1 first input vector, sum output, 16-byte aligned + * @param v2 second input vector, difference output, 16-byte aligned + * @param len length of vectors, multiple of 4 + */ + void (*butterflies_fixed)(int *av_restrict v1, int *av_restrict v2, int len); +} AVFixedDSPContext; + +/** + * Allocate and initialize a fixed DSP context. + * note: should be freed with a av_free call when no longer needed. + * + * @param strict setting to non-zero avoids using functions which may not be IEEE-754 compliant + */ +AVFixedDSPContext * avpriv_alloc_fixed_dsp(int strict); + +void ff_fixed_dsp_init_x86(AVFixedDSPContext *fdsp); + +/** + * Calculate the square root + * + * @param x input fixed point number + * + * @param bits format of fixed point number (32 - bits).bits + * + * note: input is normalized to (0, 1) fixed point value + */ + +static av_always_inline int fixed_sqrt(int x, int bits) +{ + int retval, bit_mask, guess, square, i; + int64_t accu; + int shift1 = 30 - bits; + int shift2 = bits - 15; + + if (shift1 > 0) retval = ff_sqrt(x << shift1); + else retval = ff_sqrt(x >> -shift1); + + if (shift2 > 0) { + retval = retval << shift2; + bit_mask = (1 << (shift2 - 1)); + + for (i=0; i<shift2; i++){ + guess = retval + bit_mask; + accu = (int64_t)guess * guess; + square = (int)((accu + bit_mask) >> bits); + if (x >= square) + retval += bit_mask; + bit_mask >>= 1; + } + + } + else retval >>= (-shift2); + + return retval; +} + +#endif /* AVUTIL_FIXED_DSP_H */ diff --git a/libavutil/float_dsp-test.c b/libavutil/float_dsp-test.c index e24018c07d..2603ee565d 100644 --- a/libavutil/float_dsp-test.c +++ b/libavutil/float_dsp-test.c @@ -1,26 +1,34 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ #include <float.h> +#include <math.h> #include <stdint.h> #include <stdlib.h> #include <string.h> - +#if HAVE_UNISTD_H +#include <unistd.h> /* for getopt */ +#endif +#if !HAVE_GETOPT +#include "compat/getopt.c" +#endif + +#include "common.h" #include "cpu.h" #include "internal.h" #include "lfg.h" @@ -242,9 +250,9 @@ static int test_scalarproduct_float(AVFloatDSPContext *fdsp, AVFloatDSPContext * int main(int argc, char **argv) { - int ret = 0; + int ret = 0, seeded = 0; uint32_t seed; - AVFloatDSPContext fdsp, cdsp; + AVFloatDSPContext *fdsp, *cdsp; AVLFG lfg; LOCAL_ALIGNED(32, float, src0, [LEN]); @@ -253,12 +261,40 @@ int main(int argc, char **argv) LOCAL_ALIGNED(32, double, dbl_src0, [LEN]); LOCAL_ALIGNED(32, double, dbl_src1, [LEN]); - if (argc > 2 && !strcmp(argv[1], "-s")) - seed = strtoul(argv[2], NULL, 10); - else + for (;;) { + int arg = getopt(argc, argv, "s:c:"); + if (arg == -1) + break; + switch (arg) { + case 's': + seed = strtoul(optarg, NULL, 10); + seeded = 1; + break; + case 'c': + { + int cpuflags = av_get_cpu_flags(); + + if (av_parse_cpu_caps(&cpuflags, optarg) < 0) + return 1; + + av_force_cpu_flags(cpuflags); + break; + } + } + } + if (!seeded) seed = av_get_random_seed(); - av_log(NULL, AV_LOG_INFO, "float_dsp-test: random seed %u\n", seed); + av_log(NULL, AV_LOG_INFO, "float_dsp-test: %s %u\n", seeded ? "seed" : "random seed", seed); + + fdsp = avpriv_float_dsp_alloc(1); + av_force_cpu_flags(0); + cdsp = avpriv_float_dsp_alloc(1); + + if (!fdsp || !cdsp) { + ret = 1; + goto end; + } av_lfg_init(&lfg, seed); @@ -269,28 +305,27 @@ int main(int argc, char **argv) fill_double_array(&lfg, dbl_src0, LEN); fill_double_array(&lfg, dbl_src1, LEN); - avpriv_float_dsp_init(&fdsp, 1); - av_set_cpu_flags_mask(0); - avpriv_float_dsp_init(&cdsp, 1); - - if (test_vector_fmul(&fdsp, &cdsp, src0, src1)) + if (test_vector_fmul(fdsp, cdsp, src0, src1)) ret -= 1 << 0; - if (test_vector_fmac_scalar(&fdsp, &cdsp, src2, src0, src1[0])) + if (test_vector_fmac_scalar(fdsp, cdsp, src2, src0, src1[0])) ret -= 1 << 1; - if (test_vector_fmul_scalar(&fdsp, &cdsp, src0, src1[0])) + if (test_vector_fmul_scalar(fdsp, cdsp, src0, src1[0])) ret -= 1 << 2; - if (test_vector_fmul_window(&fdsp, &cdsp, src0, src1, src2)) + if (test_vector_fmul_window(fdsp, cdsp, src0, src1, src2)) ret -= 1 << 3; - if (test_vector_fmul_add(&fdsp, &cdsp, src0, src1, src2)) + if (test_vector_fmul_add(fdsp, cdsp, src0, src1, src2)) ret -= 1 << 4; - if (test_vector_fmul_reverse(&fdsp, &cdsp, src0, src1)) + if (test_vector_fmul_reverse(fdsp, cdsp, src0, src1)) ret -= 1 << 5; - if (test_butterflies_float(&fdsp, &cdsp, src0, src1)) + if (test_butterflies_float(fdsp, cdsp, src0, src1)) ret -= 1 << 6; - if (test_scalarproduct_float(&fdsp, &cdsp, src0, src1)) + if (test_scalarproduct_float(fdsp, cdsp, src0, src1)) ret -= 1 << 7; - if (test_vector_dmul_scalar(&fdsp, &cdsp, dbl_src0, dbl_src1[0])) + if (test_vector_dmul_scalar(fdsp, cdsp, dbl_src0, dbl_src1[0])) ret -= 1 << 8; +end: + av_freep(&fdsp); + av_freep(&cdsp); return ret; } diff --git a/libavutil/float_dsp.c b/libavutil/float_dsp.c index d31fc3d166..c85daffc6a 100644 --- a/libavutil/float_dsp.c +++ b/libavutil/float_dsp.c @@ -1,24 +1,28 @@ /* - * This file is part of Libav. + * Copyright 2005 Balatoni Denes + * Copyright 2006 Loren Merritt * - * Libav is free software; you can redistribute it and/or + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ #include "config.h" #include "attributes.h" #include "float_dsp.h" +#include "mem.h" static void vector_fmul_c(float *dst, const float *src0, const float *src1, int len) @@ -89,7 +93,7 @@ static void vector_fmul_reverse_c(float *dst, const float *src0, dst[i] = src0[i] * src1[-i]; } -static void butterflies_float_c(float *restrict v1, float *restrict v2, +static void butterflies_float_c(float *av_restrict v1, float *av_restrict v2, int len) { int i; @@ -112,8 +116,12 @@ float avpriv_scalarproduct_float_c(const float *v1, const float *v2, int len) return p; } -av_cold void avpriv_float_dsp_init(AVFloatDSPContext *fdsp, int bit_exact) +av_cold AVFloatDSPContext *avpriv_float_dsp_alloc(int bit_exact) { + AVFloatDSPContext *fdsp = av_mallocz(sizeof(AVFloatDSPContext)); + if (!fdsp) + return NULL; + fdsp->vector_fmul = vector_fmul_c; fdsp->vector_fmac_scalar = vector_fmac_scalar_c; fdsp->vector_fmul_scalar = vector_fmul_scalar_c; @@ -132,4 +140,7 @@ av_cold void avpriv_float_dsp_init(AVFloatDSPContext *fdsp, int bit_exact) ff_float_dsp_init_ppc(fdsp, bit_exact); if (ARCH_X86) ff_float_dsp_init_x86(fdsp); + if (ARCH_MIPS) + ff_float_dsp_init_mips(fdsp); + return fdsp; } diff --git a/libavutil/float_dsp.h b/libavutil/float_dsp.h index 3142df4fd3..d1be38f947 100644 --- a/libavutil/float_dsp.h +++ b/libavutil/float_dsp.h @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -23,7 +23,7 @@ typedef struct AVFloatDSPContext { /** - * Calculate the product of two vectors of floats and store the result in + * Calculate the entry wise product of two vectors of floats and store the result in * a vector of floats. * * @param dst output vector @@ -104,7 +104,7 @@ typedef struct AVFloatDSPContext { const float *src1, const float *win, int len); /** - * Calculate the product of two vectors of floats, add a third vector of + * Calculate the entry wise product of two vectors of floats, add a third vector of * floats and store the result in a vector of floats. * * @param dst output vector @@ -122,7 +122,7 @@ typedef struct AVFloatDSPContext { const float *src2, int len); /** - * Calculate the product of two vectors of floats, and store the result + * Calculate the entry wise product of two vectors of floats, and store the result * in a vector of floats. The second vector of floats is iterated over * in reverse order. * @@ -145,7 +145,7 @@ typedef struct AVFloatDSPContext { * @param v2 second input vector, difference output, 16-byte aligned * @param len length of vectors, multiple of 4 */ - void (*butterflies_float)(float *restrict v1, float *restrict v2, int len); + void (*butterflies_float)(float *av_restrict v1, float *av_restrict v2, int len); /** * Calculate the scalar product of two vectors of floats. @@ -170,18 +170,17 @@ typedef struct AVFloatDSPContext { */ float avpriv_scalarproduct_float_c(const float *v1, const float *v2, int len); -/** - * Initialize a float DSP context. - * - * @param fdsp float DSP context - * @param strict setting to non-zero avoids using functions which may not be IEEE-754 compliant - */ -void avpriv_float_dsp_init(AVFloatDSPContext *fdsp, int strict); - - void ff_float_dsp_init_aarch64(AVFloatDSPContext *fdsp); void ff_float_dsp_init_arm(AVFloatDSPContext *fdsp); void ff_float_dsp_init_ppc(AVFloatDSPContext *fdsp, int strict); void ff_float_dsp_init_x86(AVFloatDSPContext *fdsp); +void ff_float_dsp_init_mips(AVFloatDSPContext *fdsp); + +/** + * Allocate a float DSP context. + * + * @param strict setting to non-zero avoids using functions which may not be IEEE-754 compliant + */ +AVFloatDSPContext *avpriv_float_dsp_alloc(int strict); #endif /* AVUTIL_FLOAT_DSP_H */ diff --git a/libavutil/frame.c b/libavutil/frame.c index ae80cee062..5607206a09 100644 --- a/libavutil/frame.c +++ b/libavutil/frame.c @@ -1,22 +1,23 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ #include "channel_layout.h" +#include "avassert.h" #include "buffer.h" #include "common.h" #include "dict.h" @@ -25,6 +26,71 @@ #include "mem.h" #include "samplefmt.h" +MAKE_ACCESSORS(AVFrame, frame, int64_t, best_effort_timestamp) +MAKE_ACCESSORS(AVFrame, frame, int64_t, pkt_duration) +MAKE_ACCESSORS(AVFrame, frame, int64_t, pkt_pos) +MAKE_ACCESSORS(AVFrame, frame, int64_t, channel_layout) +MAKE_ACCESSORS(AVFrame, frame, int, channels) +MAKE_ACCESSORS(AVFrame, frame, int, sample_rate) +MAKE_ACCESSORS(AVFrame, frame, AVDictionary *, metadata) +MAKE_ACCESSORS(AVFrame, frame, int, decode_error_flags) +MAKE_ACCESSORS(AVFrame, frame, int, pkt_size) +MAKE_ACCESSORS(AVFrame, frame, enum AVColorSpace, colorspace) +MAKE_ACCESSORS(AVFrame, frame, enum AVColorRange, color_range) + +#define CHECK_CHANNELS_CONSISTENCY(frame) \ + av_assert2(!(frame)->channel_layout || \ + (frame)->channels == \ + av_get_channel_layout_nb_channels((frame)->channel_layout)) + +AVDictionary **avpriv_frame_get_metadatap(AVFrame *frame) {return &frame->metadata;}; + +#if FF_API_FRAME_QP +int av_frame_set_qp_table(AVFrame *f, AVBufferRef *buf, int stride, int qp_type) +{ + av_buffer_unref(&f->qp_table_buf); + + f->qp_table_buf = buf; + +FF_DISABLE_DEPRECATION_WARNINGS + f->qscale_table = buf->data; + f->qstride = stride; + f->qscale_type = qp_type; +FF_ENABLE_DEPRECATION_WARNINGS + + return 0; +} + +int8_t *av_frame_get_qp_table(AVFrame *f, int *stride, int *type) +{ +FF_DISABLE_DEPRECATION_WARNINGS + *stride = f->qstride; + *type = f->qscale_type; +FF_ENABLE_DEPRECATION_WARNINGS + + if (!f->qp_table_buf) + return NULL; + + return f->qp_table_buf->data; +} +#endif + +const char *av_get_colorspace_name(enum AVColorSpace val) +{ + static const char * const name[] = { + [AVCOL_SPC_RGB] = "GBR", + [AVCOL_SPC_BT709] = "bt709", + [AVCOL_SPC_FCC] = "fcc", + [AVCOL_SPC_BT470BG] = "bt470bg", + [AVCOL_SPC_SMPTE170M] = "smpte170m", + [AVCOL_SPC_SMPTE240M] = "smpte240m", + [AVCOL_SPC_YCOCG] = "YCgCo", + }; + if ((unsigned)val >= FF_ARRAY_ELEMS(name)) + return NULL; + return name[val]; +} + static void get_frame_defaults(AVFrame *frame) { if (frame->extended_data != frame->data) @@ -32,7 +98,13 @@ static void get_frame_defaults(AVFrame *frame) memset(frame, 0, sizeof(*frame)); - frame->pts = AV_NOPTS_VALUE; + frame->pts = + frame->pkt_dts = + frame->pkt_pts = AV_NOPTS_VALUE; + av_frame_set_best_effort_timestamp(frame, AV_NOPTS_VALUE); + av_frame_set_pkt_duration (frame, 0); + av_frame_set_pkt_pos (frame, -1); + av_frame_set_pkt_size (frame, -1); frame->key_frame = 1; frame->sample_aspect_ratio = (AVRational){ 0, 1 }; frame->format = -1; /* unknown */ @@ -48,7 +120,7 @@ static void free_side_data(AVFrameSideData **ptr_sd) { AVFrameSideData *sd = *ptr_sd; - av_freep(&sd->data); + av_buffer_unref(&sd->buf); av_dict_free(&sd->metadata); av_freep(ptr_sd); } @@ -72,6 +144,7 @@ AVFrame *av_frame_alloc(void) if (!frame) return NULL; + frame->extended_data = NULL; get_frame_defaults(frame); return frame; @@ -98,21 +171,25 @@ static int get_video_buffer(AVFrame *frame, int align) return ret; if (!frame->linesize[0]) { - ret = av_image_fill_linesizes(frame->linesize, frame->format, - frame->width); - if (ret < 0) - return ret; + for(i=1; i<=align; i+=i) { + ret = av_image_fill_linesizes(frame->linesize, frame->format, + FFALIGN(frame->width, i)); + if (ret < 0) + return ret; + if (!(frame->linesize[0] & (align-1))) + break; + } for (i = 0; i < 4 && frame->linesize[i]; i++) frame->linesize[i] = FFALIGN(frame->linesize[i], align); } for (i = 0; i < 4 && frame->linesize[i]; i++) { - int h = frame->height; + int h = FFALIGN(frame->height, 32); if (i == 1 || i == 2) h = AV_CEIL_RSHIFT(h, desc->log2_chroma_h); - frame->buf[i] = av_buffer_alloc(frame->linesize[i] * h); + frame->buf[i] = av_buffer_alloc(frame->linesize[i] * h + 16 + 16/*STRIDE_ALIGN*/ - 1); if (!frame->buf[i]) goto fail; @@ -120,7 +197,7 @@ static int get_video_buffer(AVFrame *frame, int align) } if (desc->flags & AV_PIX_FMT_FLAG_PAL || desc->flags & AV_PIX_FMT_FLAG_PSEUDOPAL) { av_buffer_unref(&frame->buf[1]); - frame->buf[1] = av_buffer_alloc(1024); + frame->buf[1] = av_buffer_alloc(AVPALETTE_SIZE); if (!frame->buf[1]) goto fail; frame->data[1] = frame->buf[1]->data; @@ -136,11 +213,18 @@ fail: static int get_audio_buffer(AVFrame *frame, int align) { - int channels = av_get_channel_layout_nb_channels(frame->channel_layout); + int channels; int planar = av_sample_fmt_is_planar(frame->format); - int planes = planar ? channels : 1; + int planes; int ret, i; + if (!frame->channels) + frame->channels = av_get_channel_layout_nb_channels(frame->channel_layout); + + channels = frame->channels; + planes = planar ? channels : 1; + + CHECK_CHANNELS_CONSISTENCY(frame); if (!frame->linesize[0]) { ret = av_samples_get_buffer_size(&frame->linesize[0], channels, frame->nb_samples, frame->format, @@ -150,9 +234,9 @@ static int get_audio_buffer(AVFrame *frame, int align) } if (planes > AV_NUM_DATA_POINTERS) { - frame->extended_data = av_mallocz(planes * + frame->extended_data = av_mallocz_array(planes, sizeof(*frame->extended_data)); - frame->extended_buf = av_mallocz((planes - AV_NUM_DATA_POINTERS) * + frame->extended_buf = av_mallocz_array((planes - AV_NUM_DATA_POINTERS), sizeof(*frame->extended_buf)); if (!frame->extended_data || !frame->extended_buf) { av_freep(&frame->extended_data); @@ -190,12 +274,103 @@ int av_frame_get_buffer(AVFrame *frame, int align) if (frame->width > 0 && frame->height > 0) return get_video_buffer(frame, align); - else if (frame->nb_samples > 0 && frame->channel_layout) + else if (frame->nb_samples > 0 && (frame->channel_layout || frame->channels > 0)) return get_audio_buffer(frame, align); return AVERROR(EINVAL); } +static int frame_copy_props(AVFrame *dst, const AVFrame *src, int force_copy) +{ + int i; + + dst->key_frame = src->key_frame; + dst->pict_type = src->pict_type; + dst->sample_aspect_ratio = src->sample_aspect_ratio; + dst->pts = src->pts; + dst->repeat_pict = src->repeat_pict; + dst->interlaced_frame = src->interlaced_frame; + dst->top_field_first = src->top_field_first; + dst->palette_has_changed = src->palette_has_changed; + dst->sample_rate = src->sample_rate; + dst->opaque = src->opaque; + dst->pkt_pts = src->pkt_pts; + dst->pkt_dts = src->pkt_dts; + dst->pkt_pos = src->pkt_pos; + dst->pkt_size = src->pkt_size; + dst->pkt_duration = src->pkt_duration; + dst->reordered_opaque = src->reordered_opaque; + dst->quality = src->quality; + dst->best_effort_timestamp = src->best_effort_timestamp; + dst->coded_picture_number = src->coded_picture_number; + dst->display_picture_number = src->display_picture_number; + dst->flags = src->flags; + dst->decode_error_flags = src->decode_error_flags; + dst->color_primaries = src->color_primaries; + dst->color_trc = src->color_trc; + dst->colorspace = src->colorspace; + dst->color_range = src->color_range; + dst->chroma_location = src->chroma_location; + + av_dict_copy(&dst->metadata, src->metadata, 0); + +#if FF_API_ERROR_FRAME +FF_DISABLE_DEPRECATION_WARNINGS + memcpy(dst->error, src->error, sizeof(dst->error)); +FF_ENABLE_DEPRECATION_WARNINGS +#endif + + for (i = 0; i < src->nb_side_data; i++) { + const AVFrameSideData *sd_src = src->side_data[i]; + AVFrameSideData *sd_dst; + if ( sd_src->type == AV_FRAME_DATA_PANSCAN + && (src->width != dst->width || src->height != dst->height)) + continue; + if (force_copy) { + sd_dst = av_frame_new_side_data(dst, sd_src->type, + sd_src->size); + if (!sd_dst) { + wipe_side_data(dst); + return AVERROR(ENOMEM); + } + memcpy(sd_dst->data, sd_src->data, sd_src->size); + } else { + sd_dst = av_frame_new_side_data(dst, sd_src->type, 0); + if (!sd_dst) { + wipe_side_data(dst); + return AVERROR(ENOMEM); + } + sd_dst->buf = av_buffer_ref(sd_src->buf); + if (!sd_dst->buf) { + wipe_side_data(dst); + return AVERROR(ENOMEM); + } + sd_dst->data = sd_dst->buf->data; + sd_dst->size = sd_dst->buf->size; + } + av_dict_copy(&sd_dst->metadata, sd_src->metadata, 0); + } + +#if FF_API_FRAME_QP +FF_DISABLE_DEPRECATION_WARNINGS + dst->qscale_table = NULL; + dst->qstride = 0; + dst->qscale_type = 0; + av_buffer_unref(&dst->qp_table_buf); + if (src->qp_table_buf) { + dst->qp_table_buf = av_buffer_ref(src->qp_table_buf); + if (dst->qp_table_buf) { + dst->qscale_table = dst->qp_table_buf->data; + dst->qstride = src->qstride; + dst->qscale_type = src->qscale_type; + } + } +FF_ENABLE_DEPRECATION_WARNINGS +#endif + + return 0; +} + int av_frame_ref(AVFrame *dst, const AVFrame *src) { int i, ret = 0; @@ -203,10 +378,11 @@ int av_frame_ref(AVFrame *dst, const AVFrame *src) dst->format = src->format; dst->width = src->width; dst->height = src->height; + dst->channels = src->channels; dst->channel_layout = src->channel_layout; dst->nb_samples = src->nb_samples; - ret = av_frame_copy_props(dst, src); + ret = frame_copy_props(dst, src, 0); if (ret < 0) return ret; @@ -224,7 +400,9 @@ int av_frame_ref(AVFrame *dst, const AVFrame *src) } /* ref the buffers */ - for (i = 0; i < FF_ARRAY_ELEMS(src->buf) && src->buf[i]; i++) { + for (i = 0; i < FF_ARRAY_ELEMS(src->buf); i++) { + if (!src->buf[i]) + continue; dst->buf[i] = av_buffer_ref(src->buf[i]); if (!dst->buf[i]) { ret = AVERROR(ENOMEM); @@ -233,7 +411,7 @@ int av_frame_ref(AVFrame *dst, const AVFrame *src) } if (src->extended_buf) { - dst->extended_buf = av_mallocz(sizeof(*dst->extended_buf) * + dst->extended_buf = av_mallocz_array(sizeof(*dst->extended_buf), src->nb_extended_buf); if (!dst->extended_buf) { ret = AVERROR(ENOMEM); @@ -260,14 +438,15 @@ int av_frame_ref(AVFrame *dst, const AVFrame *src) /* duplicate extended data */ if (src->extended_data != src->data) { - int ch = av_get_channel_layout_nb_channels(src->channel_layout); + int ch = src->channels; if (!ch) { ret = AVERROR(EINVAL); goto fail; } + CHECK_CHANNELS_CONSISTENCY(src); - dst->extended_data = av_malloc(sizeof(*dst->extended_data) * ch); + dst->extended_data = av_malloc_array(sizeof(*dst->extended_data), ch); if (!dst->extended_data) { ret = AVERROR(ENOMEM); goto fail; @@ -303,6 +482,9 @@ void av_frame_unref(AVFrame *frame) { int i; + if (!frame) + return; + wipe_side_data(frame); for (i = 0; i < FF_ARRAY_ELEMS(frame->buf); i++) @@ -310,6 +492,10 @@ void av_frame_unref(AVFrame *frame) for (i = 0; i < frame->nb_extended_buf; i++) av_buffer_unref(&frame->extended_buf[i]); av_freep(&frame->extended_buf); + av_dict_free(&frame->metadata); +#if FF_API_FRAME_QP + av_buffer_unref(&frame->qp_table_buf); +#endif av_buffer_unref(&frame->hw_frames_ctx); @@ -333,8 +519,9 @@ int av_frame_is_writable(AVFrame *frame) if (!frame->buf[0]) return 0; - for (i = 0; i < FF_ARRAY_ELEMS(frame->buf) && frame->buf[i]; i++) - ret &= !!av_buffer_is_writable(frame->buf[i]); + for (i = 0; i < FF_ARRAY_ELEMS(frame->buf); i++) + if (frame->buf[i]) + ret &= !!av_buffer_is_writable(frame->buf[i]); for (i = 0; i < frame->nb_extended_buf; i++) ret &= !!av_buffer_is_writable(frame->extended_buf[i]); @@ -356,6 +543,7 @@ int av_frame_make_writable(AVFrame *frame) tmp.format = frame->format; tmp.width = frame->width; tmp.height = frame->height; + tmp.channels = frame->channels; tmp.channel_layout = frame->channel_layout; tmp.nb_samples = frame->nb_samples; ret = av_frame_get_buffer(&tmp, 32); @@ -385,50 +573,7 @@ int av_frame_make_writable(AVFrame *frame) int av_frame_copy_props(AVFrame *dst, const AVFrame *src) { - int i; - - dst->key_frame = src->key_frame; - dst->pict_type = src->pict_type; - dst->sample_aspect_ratio = src->sample_aspect_ratio; - dst->pts = src->pts; - dst->repeat_pict = src->repeat_pict; - dst->interlaced_frame = src->interlaced_frame; - dst->top_field_first = src->top_field_first; - dst->palette_has_changed = src->palette_has_changed; - dst->sample_rate = src->sample_rate; - dst->opaque = src->opaque; - dst->pkt_pts = src->pkt_pts; - dst->pkt_dts = src->pkt_dts; - dst->reordered_opaque = src->reordered_opaque; - dst->quality = src->quality; - dst->coded_picture_number = src->coded_picture_number; - dst->display_picture_number = src->display_picture_number; - dst->flags = src->flags; - dst->color_primaries = src->color_primaries; - dst->color_trc = src->color_trc; - dst->colorspace = src->colorspace; - dst->color_range = src->color_range; - dst->chroma_location = src->chroma_location; - -#if FF_API_ERROR_FRAME -FF_DISABLE_DEPRECATION_WARNINGS - memcpy(dst->error, src->error, sizeof(dst->error)); -FF_ENABLE_DEPRECATION_WARNINGS -#endif - - for (i = 0; i < src->nb_side_data; i++) { - const AVFrameSideData *sd_src = src->side_data[i]; - AVFrameSideData *sd_dst = av_frame_new_side_data(dst, sd_src->type, - sd_src->size); - if (!sd_dst) { - wipe_side_data(dst); - return AVERROR(ENOMEM); - } - memcpy(sd_dst->data, sd_src->data, sd_src->size); - av_dict_copy(&sd_dst->metadata, sd_src->metadata, 0); - } - - return 0; + return frame_copy_props(dst, src, 1); } AVBufferRef *av_frame_get_plane_buffer(AVFrame *frame, int plane) @@ -437,9 +582,10 @@ AVBufferRef *av_frame_get_plane_buffer(AVFrame *frame, int plane) int planes, i; if (frame->nb_samples) { - int channels = av_get_channel_layout_nb_channels(frame->channel_layout); + int channels = frame->channels; if (!channels) return NULL; + CHECK_CHANNELS_CONSISTENCY(frame); planes = av_sample_fmt_is_planar(frame->format) ? channels : 1; } else planes = 4; @@ -480,13 +626,16 @@ AVFrameSideData *av_frame_new_side_data(AVFrame *frame, if (!ret) return NULL; - ret->data = av_malloc(size); - if (!ret->data) { - av_freep(&ret); - return NULL; - } + if (size > 0) { + ret->buf = av_buffer_alloc(size); + if (!ret->buf) { + av_freep(&ret); + return NULL; + } - ret->size = size; + ret->data = ret->buf->data; + ret->size = size; + } ret->type = type; frame->side_data[frame->nb_side_data++] = ret; @@ -511,8 +660,8 @@ static int frame_copy_video(AVFrame *dst, const AVFrame *src) const uint8_t *src_data[4]; int i, planes; - if (dst->width != src->width || - dst->height != src->height) + if (dst->width < src->width || + dst->height < src->height) return AVERROR(EINVAL); planes = av_pix_fmt_count_planes(dst->format); @@ -523,7 +672,7 @@ static int frame_copy_video(AVFrame *dst, const AVFrame *src) memcpy(src_data, src->data, sizeof(src_data)); av_image_copy(dst->data, dst->linesize, src_data, src->linesize, - dst->format, dst->width, dst->height); + dst->format, src->width, src->height); return 0; } @@ -531,14 +680,17 @@ static int frame_copy_video(AVFrame *dst, const AVFrame *src) static int frame_copy_audio(AVFrame *dst, const AVFrame *src) { int planar = av_sample_fmt_is_planar(dst->format); - int channels = av_get_channel_layout_nb_channels(dst->channel_layout); + int channels = dst->channels; int planes = planar ? channels : 1; int i; if (dst->nb_samples != src->nb_samples || + dst->channels != src->channels || dst->channel_layout != src->channel_layout) return AVERROR(EINVAL); + CHECK_CHANNELS_CONSISTENCY(src); + for (i = 0; i < planes; i++) if (!dst->extended_data[i] || !src->extended_data[i]) return AVERROR(EINVAL); @@ -575,3 +727,23 @@ void av_frame_remove_side_data(AVFrame *frame, enum AVFrameSideDataType type) } } } + +const char *av_frame_side_data_name(enum AVFrameSideDataType type) +{ + switch(type) { + case AV_FRAME_DATA_PANSCAN: return "AVPanScan"; + case AV_FRAME_DATA_A53_CC: return "ATSC A53 Part 4 Closed Captions"; + case AV_FRAME_DATA_STEREO3D: return "Stereoscopic 3d metadata"; + case AV_FRAME_DATA_MATRIXENCODING: return "AVMatrixEncoding"; + case AV_FRAME_DATA_DOWNMIX_INFO: return "Metadata relevant to a downmix procedure"; + case AV_FRAME_DATA_REPLAYGAIN: return "AVReplayGain"; + case AV_FRAME_DATA_DISPLAYMATRIX: return "3x3 displaymatrix"; + case AV_FRAME_DATA_AFD: return "Active format description"; + case AV_FRAME_DATA_MOTION_VECTORS: return "Motion vectors"; + case AV_FRAME_DATA_SKIP_SAMPLES: return "Skip samples"; + case AV_FRAME_DATA_AUDIO_SERVICE_TYPE: return "Audio service type"; + case AV_FRAME_DATA_MASTERING_DISPLAY_METADATA: return "Mastering display metadata"; + case AV_FRAME_DATA_GOP_TIMECODE: return "GOP timecode"; + } + return NULL; +} diff --git a/libavutil/frame.h b/libavutil/frame.h index a39a1ef181..2d6299b20b 100644 --- a/libavutil/frame.h +++ b/libavutil/frame.h @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -83,15 +83,44 @@ enum AVFrameSideDataType { AV_FRAME_DATA_DISPLAYMATRIX, /** * Active Format Description data consisting of a single byte as specified - * in ETSI TS 101 154 using enum AVActiveFormatDescription. + * in ETSI TS 101 154 using AVActiveFormatDescription enum. */ AV_FRAME_DATA_AFD, - + /** + * Motion vectors exported by some codecs (on demand through the export_mvs + * flag set in the libavcodec AVCodecContext flags2 option). + * The data is the AVMotionVector struct defined in + * libavutil/motion_vector.h. + */ + AV_FRAME_DATA_MOTION_VECTORS, + /** + * Recommmends skipping the specified number of samples. This is exported + * only if the "skip_manual" AVOption is set in libavcodec. + * This has the same format as AV_PKT_DATA_SKIP_SAMPLES. + * @code + * u32le number of samples to skip from start of this packet + * u32le number of samples to skip from end of this packet + * u8 reason for start skip + * u8 reason for end skip (0=padding silence, 1=convergence) + * @endcode + */ + AV_FRAME_DATA_SKIP_SAMPLES, /** * This side data must be associated with an audio frame and corresponds to * enum AVAudioServiceType defined in avcodec.h. */ AV_FRAME_DATA_AUDIO_SERVICE_TYPE, + /** + * Mastering display metadata associated with a video frame. The payload is + * an AVMasteringDisplayMetadata type and contains information about the + * mastering display color volume. + */ + AV_FRAME_DATA_MASTERING_DISPLAY_METADATA, + /** + * The GOP timecode in 25 bit timecode format. Data format is 64-bit integer. + * This is set on the first frame of a GOP that has a temporal reference of 0. + */ + AV_FRAME_DATA_GOP_TIMECODE }; enum AVActiveFormatDescription { @@ -104,11 +133,19 @@ enum AVActiveFormatDescription { AV_AFD_SP_4_3 = 15, }; + +/** + * Structure to hold side data for an AVFrame. + * + * sizeof(AVFrameSideData) is not a part of the public ABI, so new fields may be added + * to the end with a minor bump. + */ typedef struct AVFrameSideData { enum AVFrameSideDataType type; uint8_t *data; int size; AVDictionary *metadata; + AVBufferRef *buf; } AVFrameSideData; /** @@ -136,12 +173,23 @@ typedef struct AVFrameSideData { * * sizeof(AVFrame) is not a part of the public ABI, so new fields may be added * to the end with a minor bump. + * Similarly fields that are marked as to be only accessed by + * av_opt_ptr() can be reordered. This allows 2 forks to add fields + * without breaking compatibility with each other. */ typedef struct AVFrame { #define AV_NUM_DATA_POINTERS 8 /** * pointer to the picture/channel planes. * This might be different from the first allocated byte + * + * Some decoders access areas outside 0,0 - width,height, please + * see avcodec_align_dimensions2(). Some filters and swscale can read + * up to 16 bytes beyond the planes, if these filters are to be used, + * then 16 extra bytes must be allocated. + * + * NOTE: Except for hwaccel formats, pointers not needed by the format + * MUST be set to NULL. */ uint8_t *data[AV_NUM_DATA_POINTERS]; @@ -152,6 +200,11 @@ typedef struct AVFrame { * For audio, only linesize[0] may be set. For planar audio, each channel * plane must be the same size. * + * For video the linesizes should be multiples of the CPUs alignment + * preference, this is 16 or 32 for modern desktop CPUs. + * Some code requires such alignment other code can be slower without + * correct alignment, for yet other it makes no difference. + * * @note The linesize may be larger than the size of usable data -- there * may be extra padding present for performance reasons. */ @@ -216,7 +269,9 @@ typedef struct AVFrame { int64_t pkt_pts; /** - * DTS copied from the AVPacket that triggered returning this frame. + * DTS copied from the AVPacket that triggered returning this frame. (if frame threading isn't used) + * This is also the Presentation time of this AVFrame calculated from + * only AVPacket.dts values without pts values. */ int64_t pkt_dts; @@ -344,12 +399,26 @@ typedef struct AVFrame { */ int flags; + /** + * MPEG vs JPEG YUV range. + * It must be accessed using av_frame_get_color_range() and + * av_frame_set_color_range(). + * - encoding: Set by user + * - decoding: Set by libavcodec + */ enum AVColorRange color_range; enum AVColorPrimaries color_primaries; enum AVColorTransferCharacteristic color_trc; + /** + * YUV colorspace type. + * It must be accessed using av_frame_get_colorspace() and + * av_frame_set_colorspace(). + * - encoding: Set by user + * - decoding: Set by libavcodec + */ enum AVColorSpace colorspace; enum AVChromaLocation chroma_location; @@ -359,9 +428,140 @@ typedef struct AVFrame { * AVHWFramesContext describing the frame. */ AVBufferRef *hw_frames_ctx; + + /** + * frame timestamp estimated using various heuristics, in stream time base + * Code outside libavutil should access this field using: + * av_frame_get_best_effort_timestamp(frame) + * - encoding: unused + * - decoding: set by libavcodec, read by user. + */ + int64_t best_effort_timestamp; + + /** + * reordered pos from the last AVPacket that has been input into the decoder + * Code outside libavutil should access this field using: + * av_frame_get_pkt_pos(frame) + * - encoding: unused + * - decoding: Read by user. + */ + int64_t pkt_pos; + + /** + * duration of the corresponding packet, expressed in + * AVStream->time_base units, 0 if unknown. + * Code outside libavutil should access this field using: + * av_frame_get_pkt_duration(frame) + * - encoding: unused + * - decoding: Read by user. + */ + int64_t pkt_duration; + + /** + * metadata. + * Code outside libavutil should access this field using: + * av_frame_get_metadata(frame) + * - encoding: Set by user. + * - decoding: Set by libavcodec. + */ + AVDictionary *metadata; + + /** + * decode error flags of the frame, set to a combination of + * FF_DECODE_ERROR_xxx flags if the decoder produced a frame, but there + * were errors during the decoding. + * Code outside libavutil should access this field using: + * av_frame_get_decode_error_flags(frame) + * - encoding: unused + * - decoding: set by libavcodec, read by user. + */ + int decode_error_flags; +#define FF_DECODE_ERROR_INVALID_BITSTREAM 1 +#define FF_DECODE_ERROR_MISSING_REFERENCE 2 + + /** + * number of audio channels, only used for audio. + * Code outside libavutil should access this field using: + * av_frame_get_channels(frame) + * - encoding: unused + * - decoding: Read by user. + */ + int channels; + + /** + * size of the corresponding packet containing the compressed + * frame. It must be accessed using av_frame_get_pkt_size() and + * av_frame_set_pkt_size(). + * It is set to a negative value if unknown. + * - encoding: unused + * - decoding: set by libavcodec, read by user. + */ + int pkt_size; + +#if FF_API_FRAME_QP + /** + * QP table + * Not to be accessed directly from outside libavutil + */ + attribute_deprecated + int8_t *qscale_table; + /** + * QP store stride + * Not to be accessed directly from outside libavutil + */ + attribute_deprecated + int qstride; + + attribute_deprecated + int qscale_type; + + /** + * Not to be accessed directly from outside libavutil + */ + AVBufferRef *qp_table_buf; +#endif } AVFrame; /** + * Accessors for some AVFrame fields. + * The position of these field in the structure is not part of the ABI, + * they should not be accessed directly outside libavutil. + */ +int64_t av_frame_get_best_effort_timestamp(const AVFrame *frame); +void av_frame_set_best_effort_timestamp(AVFrame *frame, int64_t val); +int64_t av_frame_get_pkt_duration (const AVFrame *frame); +void av_frame_set_pkt_duration (AVFrame *frame, int64_t val); +int64_t av_frame_get_pkt_pos (const AVFrame *frame); +void av_frame_set_pkt_pos (AVFrame *frame, int64_t val); +int64_t av_frame_get_channel_layout (const AVFrame *frame); +void av_frame_set_channel_layout (AVFrame *frame, int64_t val); +int av_frame_get_channels (const AVFrame *frame); +void av_frame_set_channels (AVFrame *frame, int val); +int av_frame_get_sample_rate (const AVFrame *frame); +void av_frame_set_sample_rate (AVFrame *frame, int val); +AVDictionary *av_frame_get_metadata (const AVFrame *frame); +void av_frame_set_metadata (AVFrame *frame, AVDictionary *val); +int av_frame_get_decode_error_flags (const AVFrame *frame); +void av_frame_set_decode_error_flags (AVFrame *frame, int val); +int av_frame_get_pkt_size(const AVFrame *frame); +void av_frame_set_pkt_size(AVFrame *frame, int val); +AVDictionary **avpriv_frame_get_metadatap(AVFrame *frame); +#if FF_API_FRAME_QP +int8_t *av_frame_get_qp_table(AVFrame *f, int *stride, int *type); +int av_frame_set_qp_table(AVFrame *f, AVBufferRef *buf, int stride, int type); +#endif +enum AVColorSpace av_frame_get_colorspace(const AVFrame *frame); +void av_frame_set_colorspace(AVFrame *frame, enum AVColorSpace val); +enum AVColorRange av_frame_get_color_range(const AVFrame *frame); +void av_frame_set_color_range(AVFrame *frame, enum AVColorRange val); + +/** + * Get the name of a colorspace. + * @return a static string identifying the colorspace; can be NULL. + */ +const char *av_get_colorspace_name(enum AVColorSpace val); + +/** * Allocate an AVFrame and set its fields to default values. The resulting * struct must be freed using av_frame_free(). * @@ -532,6 +732,11 @@ AVFrameSideData *av_frame_get_side_data(const AVFrame *frame, void av_frame_remove_side_data(AVFrame *frame, enum AVFrameSideDataType type); /** + * @return a string identifying the side data type + */ +const char *av_frame_side_data_name(enum AVFrameSideDataType type); + +/** * @} */ diff --git a/libavutil/hash-test.c b/libavutil/hash-test.c new file mode 100644 index 0000000000..fe14cc3b1c --- /dev/null +++ b/libavutil/hash-test.c @@ -0,0 +1,58 @@ +/* + * Copyright (C) 2013 Reimar Döffinger <Reimar.Doeffinger@gmx.de> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "hash.c" + +#define SRC_BUF_SIZE 64 +#define DST_BUF_SIZE (AV_HASH_MAX_SIZE * 8) + +int main(void) +{ + struct AVHashContext *ctx = NULL; + int i, j; + static const uint8_t src[SRC_BUF_SIZE] = { 0 }; + uint8_t dst[DST_BUF_SIZE]; + for (i = 0; i < NUM_HASHES; i++) { + if (av_hash_alloc(&ctx, av_hash_names(i)) < 0) + return 1; + + av_hash_init(ctx); + av_hash_update(ctx, src, SRC_BUF_SIZE); + memset(dst, 0, DST_BUF_SIZE); + av_hash_final_hex(ctx, dst, DST_BUF_SIZE); + printf("%s hex: %s\n", av_hash_get_name(ctx), dst); + + av_hash_init(ctx); + av_hash_update(ctx, src, SRC_BUF_SIZE); + av_hash_final_bin(ctx, dst, DST_BUF_SIZE); + printf("%s bin: ", av_hash_get_name(ctx)); + for (j = 0; j < av_hash_get_size(ctx); j++) { + printf("%#x ", dst[j]); + } + printf("\n"); + + av_hash_init(ctx); + av_hash_update(ctx, src, SRC_BUF_SIZE); + av_hash_final_b64(ctx, dst, DST_BUF_SIZE); + printf("%s b64: %s\n", av_hash_get_name(ctx), dst); + av_hash_freep(&ctx); + } + return 0; +} diff --git a/libavutil/hash.c b/libavutil/hash.c new file mode 100644 index 0000000000..7037b0d6ff --- /dev/null +++ b/libavutil/hash.c @@ -0,0 +1,239 @@ +/* + * Copyright (C) 2013 Reimar Döffinger <Reimar.Doeffinger@gmx.de> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ +#include <stdint.h> +#include "hash.h" + +#include "adler32.h" +#include "crc.h" +#include "md5.h" +#include "murmur3.h" +#include "ripemd.h" +#include "sha.h" +#include "sha512.h" + +#include "avstring.h" +#include "base64.h" +#include "error.h" +#include "intreadwrite.h" +#include "mem.h" + +enum hashtype { + MD5, + MURMUR3, + RIPEMD128, + RIPEMD160, + RIPEMD256, + RIPEMD320, + SHA160, + SHA224, + SHA256, + SHA512_224, + SHA512_256, + SHA384, + SHA512, + CRC32, + ADLER32, + NUM_HASHES +}; + +typedef struct AVHashContext { + void *ctx; + enum hashtype type; + const AVCRC *crctab; + uint32_t crc; +} AVHashContext; + +static const struct { + const char *name; + int size; +} hashdesc[] = { + [MD5] = {"MD5", 16}, + [MURMUR3] = {"murmur3", 16}, + [RIPEMD128] = {"RIPEMD128", 16}, + [RIPEMD160] = {"RIPEMD160", 20}, + [RIPEMD256] = {"RIPEMD256", 32}, + [RIPEMD320] = {"RIPEMD320", 40}, + [SHA160] = {"SHA160", 20}, + [SHA224] = {"SHA224", 28}, + [SHA256] = {"SHA256", 32}, + [SHA512_224] = {"SHA512/224", 28}, + [SHA512_256] = {"SHA512/256", 32}, + [SHA384] = {"SHA384", 48}, + [SHA512] = {"SHA512", 64}, + [CRC32] = {"CRC32", 4}, + [ADLER32] = {"adler32", 4}, +}; + +const char *av_hash_names(int i) +{ + if (i < 0 || i >= NUM_HASHES) return NULL; + return hashdesc[i].name; +} + +const char *av_hash_get_name(const AVHashContext *ctx) +{ + return hashdesc[ctx->type].name; +} + +int av_hash_get_size(const AVHashContext *ctx) +{ + return hashdesc[ctx->type].size; +} + +int av_hash_alloc(AVHashContext **ctx, const char *name) +{ + AVHashContext *res; + int i; + *ctx = NULL; + for (i = 0; i < NUM_HASHES; i++) + if (av_strcasecmp(name, hashdesc[i].name) == 0) + break; + if (i >= NUM_HASHES) return AVERROR(EINVAL); + res = av_mallocz(sizeof(*res)); + if (!res) return AVERROR(ENOMEM); + res->type = i; + switch (i) { + case MD5: res->ctx = av_md5_alloc(); break; + case MURMUR3: res->ctx = av_murmur3_alloc(); break; + case RIPEMD128: + case RIPEMD160: + case RIPEMD256: + case RIPEMD320: res->ctx = av_ripemd_alloc(); break; + case SHA160: + case SHA224: + case SHA256: res->ctx = av_sha_alloc(); break; + case SHA512_224: + case SHA512_256: + case SHA384: + case SHA512: res->ctx = av_sha512_alloc(); break; + case CRC32: res->crctab = av_crc_get_table(AV_CRC_32_IEEE_LE); break; + case ADLER32: break; + } + if (i != ADLER32 && i != CRC32 && !res->ctx) { + av_free(res); + return AVERROR(ENOMEM); + } + *ctx = res; + return 0; +} + +void av_hash_init(AVHashContext *ctx) +{ + switch (ctx->type) { + case MD5: av_md5_init(ctx->ctx); break; + case MURMUR3: av_murmur3_init(ctx->ctx); break; + case RIPEMD128: av_ripemd_init(ctx->ctx, 128); break; + case RIPEMD160: av_ripemd_init(ctx->ctx, 160); break; + case RIPEMD256: av_ripemd_init(ctx->ctx, 256); break; + case RIPEMD320: av_ripemd_init(ctx->ctx, 320); break; + case SHA160: av_sha_init(ctx->ctx, 160); break; + case SHA224: av_sha_init(ctx->ctx, 224); break; + case SHA256: av_sha_init(ctx->ctx, 256); break; + case SHA512_224: av_sha512_init(ctx->ctx, 224); break; + case SHA512_256: av_sha512_init(ctx->ctx, 256); break; + case SHA384: av_sha512_init(ctx->ctx, 384); break; + case SHA512: av_sha512_init(ctx->ctx, 512); break; + case CRC32: ctx->crc = UINT32_MAX; break; + case ADLER32: ctx->crc = 1; break; + } +} + +void av_hash_update(AVHashContext *ctx, const uint8_t *src, int len) +{ + switch (ctx->type) { + case MD5: av_md5_update(ctx->ctx, src, len); break; + case MURMUR3: av_murmur3_update(ctx->ctx, src, len); break; + case RIPEMD128: + case RIPEMD160: + case RIPEMD256: + case RIPEMD320: av_ripemd_update(ctx->ctx, src, len); break; + case SHA160: + case SHA224: + case SHA256: av_sha_update(ctx->ctx, src, len); break; + case SHA512_224: + case SHA512_256: + case SHA384: + case SHA512: av_sha512_update(ctx->ctx, src, len); break; + case CRC32: ctx->crc = av_crc(ctx->crctab, ctx->crc, src, len); break; + case ADLER32: ctx->crc = av_adler32_update(ctx->crc, src, len); break; + } +} + +void av_hash_final(AVHashContext *ctx, uint8_t *dst) +{ + switch (ctx->type) { + case MD5: av_md5_final(ctx->ctx, dst); break; + case MURMUR3: av_murmur3_final(ctx->ctx, dst); break; + case RIPEMD128: + case RIPEMD160: + case RIPEMD256: + case RIPEMD320: av_ripemd_final(ctx->ctx, dst); break; + case SHA160: + case SHA224: + case SHA256: av_sha_final(ctx->ctx, dst); break; + case SHA512_224: + case SHA512_256: + case SHA384: + case SHA512: av_sha512_final(ctx->ctx, dst); break; + case CRC32: AV_WB32(dst, ctx->crc ^ UINT32_MAX); break; + case ADLER32: AV_WB32(dst, ctx->crc); break; + } +} + +void av_hash_final_bin(struct AVHashContext *ctx, uint8_t *dst, int size) +{ + uint8_t buf[AV_HASH_MAX_SIZE]; + unsigned rsize = av_hash_get_size(ctx); + + av_hash_final(ctx, buf); + memcpy(dst, buf, FFMIN(size, rsize)); + if (size > rsize) + memset(dst + rsize, 0, size - rsize); +} + +void av_hash_final_hex(struct AVHashContext *ctx, uint8_t *dst, int size) +{ + uint8_t buf[AV_HASH_MAX_SIZE]; + unsigned rsize = av_hash_get_size(ctx), i; + + av_hash_final(ctx, buf); + for (i = 0; i < FFMIN(rsize, size / 2); i++) + snprintf(dst + i * 2, size - i * 2, "%02x", buf[i]); +} + +void av_hash_final_b64(struct AVHashContext *ctx, uint8_t *dst, int size) +{ + uint8_t buf[AV_HASH_MAX_SIZE], b64[AV_BASE64_SIZE(AV_HASH_MAX_SIZE)]; + unsigned rsize = av_hash_get_size(ctx), osize; + + av_hash_final(ctx, buf); + av_base64_encode(b64, sizeof(b64), buf, rsize); + osize = AV_BASE64_SIZE(rsize); + memcpy(dst, b64, FFMIN(osize, size)); + if (size < osize) + dst[size - 1] = 0; +} + +void av_hash_freep(AVHashContext **ctx) +{ + if (*ctx) + av_freep(&(*ctx)->ctx); + av_freep(ctx); +} diff --git a/libavutil/hash.h b/libavutil/hash.h new file mode 100644 index 0000000000..d4bcbf8cc8 --- /dev/null +++ b/libavutil/hash.h @@ -0,0 +1,112 @@ +/* + * Copyright (C) 2013 Reimar Döffinger <Reimar.Doeffinger@gmx.de> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#ifndef AVUTIL_HASH_H +#define AVUTIL_HASH_H + +#include <stdint.h> + +struct AVHashContext; + +/** + * Allocate a hash context for the algorithm specified by name. + * + * @return >= 0 for success, a negative error code for failure + * @note The context is not initialized, you must call av_hash_init(). + */ +int av_hash_alloc(struct AVHashContext **ctx, const char *name); + +/** + * Get the names of available hash algorithms. + * + * This function can be used to enumerate the algorithms. + * + * @param i index of the hash algorithm, starting from 0 + * @return a pointer to a static string or NULL if i is out of range + */ +const char *av_hash_names(int i); + +/** + * Get the name of the algorithm corresponding to the given hash context. + */ +const char *av_hash_get_name(const struct AVHashContext *ctx); + +/** + * Maximum value that av_hash_get_size will currently return. + * + * You can use this if you absolutely want or need to use static allocation + * and are fine with not supporting hashes newly added to libavutil without + * recompilation. + * Note that you still need to check against av_hash_get_size, adding new hashes + * with larger sizes will not be considered an ABI change and should not cause + * your code to overflow a buffer. + */ +#define AV_HASH_MAX_SIZE 64 + +/** + * Get the size of the resulting hash value in bytes. + * + * The pointer passed to av_hash_final have space for at least this many bytes. + */ +int av_hash_get_size(const struct AVHashContext *ctx); + +/** + * Initialize or reset a hash context. + */ +void av_hash_init(struct AVHashContext *ctx); + +/** + * Update a hash context with additional data. + */ +void av_hash_update(struct AVHashContext *ctx, const uint8_t *src, int len); + +/** + * Finalize a hash context and compute the actual hash value. + */ +void av_hash_final(struct AVHashContext *ctx, uint8_t *dst); + +/** + * Finalize a hash context and compute the actual hash value. + * If size is smaller than the hash size, the hash is truncated; + * if size is larger, the buffer is padded with 0. + */ +void av_hash_final_bin(struct AVHashContext *ctx, uint8_t *dst, int size); + +/** + * Finalize a hash context and compute the actual hash value as a hex string. + * The string is always 0-terminated. + * If size is smaller than 2 * hash_size + 1, the hex string is truncated. + */ +void av_hash_final_hex(struct AVHashContext *ctx, uint8_t *dst, int size); + +/** + * Finalize a hash context and compute the actual hash value as a base64 string. + * The string is always 0-terminated. + * If size is smaller than AV_BASE64_SIZE(hash_size), the base64 string is + * truncated. + */ +void av_hash_final_b64(struct AVHashContext *ctx, uint8_t *dst, int size); + +/** + * Free hash context. + */ +void av_hash_freep(struct AVHashContext **ctx); + +#endif /* AVUTIL_HASH_H */ diff --git a/libavutil/hmac-test.c b/libavutil/hmac-test.c index fb6bf87cff..925f3dfef6 100644 --- a/libavutil/hmac-test.c +++ b/libavutil/hmac-test.c @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -41,44 +41,52 @@ static void test(AVHMAC *hmac, const uint8_t *key, int keylen, int main(void) { uint8_t key1[20], key3[131], data3[50]; - enum AVHMACType i = AV_HMAC_SHA224; - const uint8_t key2[] = "Jefe"; - const uint8_t data1[] = "Hi There"; - const uint8_t data2[] = "what do ya want for nothing?"; - const uint8_t data4[] = "Test Using Larger Than Block-Size Key - Hash Key First"; - const uint8_t data5[] = "Test Using Larger Than Block-Size Key and Larger Than One Block-Size Data"; - const uint8_t data6[] = "This is a test using a larger than block-size key and a larger " + AVHMAC *hmac; + enum AVHMACType i; + static const uint8_t key2[] = "Jefe"; + static const uint8_t data1[] = "Hi There"; + static const uint8_t data2[] = "what do ya want for nothing?"; + static const uint8_t data4[] = "Test Using Larger Than Block-Size Key - Hash Key First"; + static const uint8_t data5[] = "Test Using Larger Than Block-Size Key and Larger Than One Block-Size Data"; + static const uint8_t data6[] = "This is a test using a larger than block-size key and a larger " "than block-size data. The key needs to be hashed before being used" " by the HMAC algorithm."; - AVHMAC *hmac = av_hmac_alloc(AV_HMAC_MD5); - if (!hmac) - return 1; memset(key1, 0x0b, sizeof(key1)); memset(key3, 0xaa, sizeof(key3)); memset(data3, 0xdd, sizeof(data3)); - // RFC 2202 test vectors - test(hmac, key1, 16, data1, sizeof(data1)); - test(hmac, key2, sizeof(key2), data2, sizeof(data2)); - test(hmac, key3, 16, data3, sizeof(data3)); - test(hmac, key3, 80, data4, sizeof(data4)); - test(hmac, key3, 80, data5, sizeof(data5)); - av_hmac_free(hmac); - /* SHA-1 */ - hmac = av_hmac_alloc(AV_HMAC_SHA1); - if (!hmac) - return 1; - // RFC 2202 test vectors - test(hmac, key1, sizeof(key1), data1, sizeof(data1)); - test(hmac, key2, sizeof(key2), data2, sizeof(data2)); - test(hmac, key3, 20, data3, sizeof(data3)); - test(hmac, key3, 80, data4, sizeof(data4)); - test(hmac, key3, 80, data5, sizeof(data5)); - av_hmac_free(hmac); + /* MD5, SHA-1 */ + for (i = AV_HMAC_MD5; i <= AV_HMAC_SHA1; i++) { + hmac = av_hmac_alloc(i); + if (!hmac) + return 1; + // RFC 2202 test vectors + test(hmac, key1, hmac->hashlen, data1, sizeof(data1)); + test(hmac, key2, sizeof(key2), data2, sizeof(data2)); + test(hmac, key3, hmac->hashlen, data3, sizeof(data3)); + test(hmac, key3, 80, data4, sizeof(data4)); + test(hmac, key3, 80, data5, sizeof(data5)); + av_hmac_free(hmac); + } /* SHA-2 */ - while (i <= AV_HMAC_SHA256) { + for (i = AV_HMAC_SHA224; i <= AV_HMAC_SHA256; i++) { + hmac = av_hmac_alloc(i); + if (!hmac) + return 1; + // RFC 4231 test vectors + test(hmac, key1, sizeof(key1), data1, sizeof(data1)); + test(hmac, key2, sizeof(key2), data2, sizeof(data2)); + test(hmac, key3, 20, data3, sizeof(data3)); + test(hmac, key3, sizeof(key3), data4, sizeof(data4)); + test(hmac, key3, sizeof(key3), data6, sizeof(data6)); + av_hmac_free(hmac); + } + + for (i = AV_HMAC_SHA384; i <= AV_HMAC_SHA512; i++) { hmac = av_hmac_alloc(i); + if (!hmac) + return 1; // RFC 4231 test vectors test(hmac, key1, sizeof(key1), data1, sizeof(data1)); test(hmac, key2, sizeof(key2), data2, sizeof(data2)); @@ -86,7 +94,6 @@ int main(void) test(hmac, key3, sizeof(key3), data4, sizeof(data4)); test(hmac, key3, sizeof(key3), data6, sizeof(data6)); av_hmac_free(hmac); - i++; } return 0; } diff --git a/libavutil/hmac.c b/libavutil/hmac.c index 378be62fd3..8ec6d70406 100644 --- a/libavutil/hmac.c +++ b/libavutil/hmac.c @@ -1,20 +1,20 @@ /* * Copyright (C) 2012 Martin Storsjo * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -24,10 +24,11 @@ #include "hmac.h" #include "md5.h" #include "sha.h" +#include "sha512.h" #include "mem.h" -#define MAX_HASHLEN 32 -#define MAX_BLOCKLEN 64 +#define MAX_HASHLEN 64 +#define MAX_BLOCKLEN 128 struct AVHMAC { void *hash; @@ -45,9 +46,17 @@ static av_cold void sha ## bits ##_init(void *ctx) \ av_sha_init(ctx, bits); \ } +#define DEFINE_SHA512(bits) \ +static av_cold void sha ## bits ##_init(void *ctx) \ +{ \ + av_sha512_init(ctx, bits); \ +} + DEFINE_SHA(160) DEFINE_SHA(224) DEFINE_SHA(256) +DEFINE_SHA512(384) +DEFINE_SHA512(512) AVHMAC *av_hmac_alloc(enum AVHMACType type) { @@ -58,35 +67,51 @@ AVHMAC *av_hmac_alloc(enum AVHMACType type) case AV_HMAC_MD5: c->blocklen = 64; c->hashlen = 16; - c->init = av_md5_init; - c->update = av_md5_update; - c->final = av_md5_final; + c->init = (void*)av_md5_init; + c->update = (void*)av_md5_update; + c->final = (void*)av_md5_final; c->hash = av_md5_alloc(); break; case AV_HMAC_SHA1: c->blocklen = 64; c->hashlen = 20; c->init = sha160_init; - c->update = av_sha_update; - c->final = av_sha_final; + c->update = (void*)av_sha_update; + c->final = (void*)av_sha_final; c->hash = av_sha_alloc(); break; case AV_HMAC_SHA224: c->blocklen = 64; c->hashlen = 28; c->init = sha224_init; - c->update = av_sha_update; - c->final = av_sha_final; + c->update = (void*)av_sha_update; + c->final = (void*)av_sha_final; c->hash = av_sha_alloc(); break; case AV_HMAC_SHA256: c->blocklen = 64; c->hashlen = 32; c->init = sha256_init; - c->update = av_sha_update; - c->final = av_sha_final; + c->update = (void*)av_sha_update; + c->final = (void*)av_sha_final; c->hash = av_sha_alloc(); break; + case AV_HMAC_SHA384: + c->blocklen = 128; + c->hashlen = 48; + c->init = sha384_init; + c->update = (void*)av_sha512_update; + c->final = (void*)av_sha512_final; + c->hash = av_sha512_alloc(); + break; + case AV_HMAC_SHA512: + c->blocklen = 128; + c->hashlen = 64; + c->init = sha512_init; + c->update = (void*)av_sha512_update; + c->final = (void*)av_sha512_final; + c->hash = av_sha512_alloc(); + break; default: av_free(c); return NULL; @@ -102,7 +127,7 @@ void av_hmac_free(AVHMAC *c) { if (!c) return; - av_free(c->hash); + av_freep(&c->hash); av_free(c); } diff --git a/libavutil/hmac.h b/libavutil/hmac.h index 1ed56269cb..576a0a4fb9 100644 --- a/libavutil/hmac.h +++ b/libavutil/hmac.h @@ -1,20 +1,20 @@ /* * Copyright (C) 2012 Martin Storsjo * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -23,6 +23,7 @@ #include <stdint.h> +#include "version.h" /** * @defgroup lavu_hmac HMAC * @ingroup lavu_crypto @@ -34,6 +35,8 @@ enum AVHMACType { AV_HMAC_SHA1, AV_HMAC_SHA224, AV_HMAC_SHA256, + AV_HMAC_SHA384 = 12, + AV_HMAC_SHA512, }; typedef struct AVHMAC AVHMAC; diff --git a/libavutil/hwcontext.c b/libavutil/hwcontext.c index ac1e2c9d67..b34b317d5d 100644 --- a/libavutil/hwcontext.c +++ b/libavutil/hwcontext.c @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -173,7 +173,7 @@ AVBufferRef *av_hwframe_ctx_alloc(AVBufferRef *device_ref_in) AVHWDeviceContext *device_ctx = (AVHWDeviceContext*)device_ref_in->data; const HWContextType *hw_type = device_ctx->internal->hw_type; AVHWFramesContext *ctx; - AVBufferRef *buf, *device_ref = NULL;; + AVBufferRef *buf, *device_ref = NULL; ctx = av_mallocz(sizeof(*ctx)); if (!ctx) diff --git a/libavutil/hwcontext.h b/libavutil/hwcontext.h index 8502342409..44be1977a8 100644 --- a/libavutil/hwcontext.h +++ b/libavutil/hwcontext.h @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/hwcontext_cuda.c b/libavutil/hwcontext_cuda.c index 6b87b61175..2c5980d619 100644 --- a/libavutil/hwcontext_cuda.c +++ b/libavutil/hwcontext_cuda.c @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/hwcontext_cuda.h b/libavutil/hwcontext_cuda.h index 7f067c7572..23a77cee73 100644 --- a/libavutil/hwcontext_cuda.h +++ b/libavutil/hwcontext_cuda.h @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/hwcontext_internal.h b/libavutil/hwcontext_internal.h index b7828c0350..f18b616fca 100644 --- a/libavutil/hwcontext_internal.h +++ b/libavutil/hwcontext_internal.h @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/hwcontext_vaapi.c b/libavutil/hwcontext_vaapi.c index 1a385ba514..c2cdaa95c3 100644 --- a/libavutil/hwcontext_vaapi.c +++ b/libavutil/hwcontext_vaapi.c @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/hwcontext_vaapi.h b/libavutil/hwcontext_vaapi.h index 1c87f5db5e..7fd1a36e8f 100644 --- a/libavutil/hwcontext_vaapi.h +++ b/libavutil/hwcontext_vaapi.h @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/hwcontext_vdpau.c b/libavutil/hwcontext_vdpau.c index faae5f8641..37b740eaad 100644 --- a/libavutil/hwcontext_vdpau.c +++ b/libavutil/hwcontext_vdpau.c @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/hwcontext_vdpau.h b/libavutil/hwcontext_vdpau.h index 358e942fca..1b7ea1e443 100644 --- a/libavutil/hwcontext_vdpau.h +++ b/libavutil/hwcontext_vdpau.h @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/imgutils.c b/libavutil/imgutils.c index 6c581396a3..37808e53d0 100644 --- a/libavutil/imgutils.c +++ b/libavutil/imgutils.c @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -21,6 +21,7 @@ * misc image utilities */ +#include "avassert.h" #include "common.h" #include "imgutils.h" #include "internal.h" @@ -48,27 +49,45 @@ void av_image_fill_max_pixsteps(int max_pixsteps[4], int max_pixstep_comps[4], } } +static inline +int image_get_linesize(int width, int plane, + int max_step, int max_step_comp, + const AVPixFmtDescriptor *desc) +{ + int s, shifted_w, linesize; + + if (!desc) + return AVERROR(EINVAL); + + if (width < 0) + return AVERROR(EINVAL); + s = (max_step_comp == 1 || max_step_comp == 2) ? desc->log2_chroma_w : 0; + shifted_w = ((width + (1 << s) - 1)) >> s; + if (shifted_w && max_step > INT_MAX / shifted_w) + return AVERROR(EINVAL); + linesize = max_step * shifted_w; + + if (desc->flags & AV_PIX_FMT_FLAG_BITSTREAM) + linesize = (linesize + 7) >> 3; + return linesize; +} + int av_image_get_linesize(enum AVPixelFormat pix_fmt, int width, int plane) { const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt); int max_step [4]; /* max pixel step for each plane */ int max_step_comp[4]; /* the component for each plane which has the max pixel step */ - int s; - if (!desc) + if (!desc || desc->flags & AV_PIX_FMT_FLAG_HWACCEL) return AVERROR(EINVAL); - if (desc->flags & AV_PIX_FMT_FLAG_BITSTREAM) - return (width * desc->comp[0].step + 7) >> 3; - av_image_fill_max_pixsteps(max_step, max_step_comp, desc); - s = (max_step_comp[plane] == 1 || max_step_comp[plane] == 2) ? desc->log2_chroma_w : 0; - return max_step[plane] * (((width + (1 << s) - 1)) >> s); + return image_get_linesize(width, plane, max_step[plane], max_step_comp[plane], desc); } int av_image_fill_linesizes(int linesizes[4], enum AVPixelFormat pix_fmt, int width) { - int i; + int i, ret; const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt); int max_step [4]; /* max pixel step for each plane */ int max_step_comp[4]; /* the component for each plane which has the max pixel step */ @@ -78,20 +97,11 @@ int av_image_fill_linesizes(int linesizes[4], enum AVPixelFormat pix_fmt, int wi if (!desc || desc->flags & AV_PIX_FMT_FLAG_HWACCEL) return AVERROR(EINVAL); - if (desc->flags & AV_PIX_FMT_FLAG_BITSTREAM) { - if (width > (INT_MAX - 7) / desc->comp[0].step) - return AVERROR(EINVAL); - linesizes[0] = (width * desc->comp[0].step + 7) >> 3; - return 0; - } - av_image_fill_max_pixsteps(max_step, max_step_comp, desc); for (i = 0; i < 4; i++) { - int s = (max_step_comp[i] == 1 || max_step_comp[i] == 2) ? desc->log2_chroma_w : 0; - int shifted_w = ((width + (1 << s) - 1)) >> s; - if (max_step[i] > INT_MAX / shifted_w) - return AVERROR(EINVAL); - linesizes[i] = max_step[i] * shifted_w; + if ((ret = image_get_linesize(width, i, max_step[i], max_step_comp[i], desc)) < 0) + return ret; + linesizes[i] = ret; } return 0; @@ -115,7 +125,6 @@ int av_image_fill_pointers(uint8_t *data[4], enum AVPixelFormat pix_fmt, int hei if (desc->flags & AV_PIX_FMT_FLAG_PAL || desc->flags & AV_PIX_FMT_FLAG_PSEUDOPAL) { - size[0] = (size[0] + 3) & ~3; data[1] = ptr + size[0]; /* palette is stored here as 256 32 bits words */ return size[0] + 256 * 4; } @@ -191,7 +200,7 @@ int av_image_alloc(uint8_t *pointers[4], int linesizes[4], if ((ret = av_image_check_size(w, h, 0, NULL)) < 0) return ret; - if ((ret = av_image_fill_linesizes(linesizes, pix_fmt, w)) < 0) + if ((ret = av_image_fill_linesizes(linesizes, pix_fmt, align>7 ? FFALIGN(w, 8) : w)) < 0) return ret; for (i = 0; i < 4; i++) @@ -206,8 +215,21 @@ int av_image_alloc(uint8_t *pointers[4], int linesizes[4], av_free(buf); return ret; } - if (desc->flags & AV_PIX_FMT_FLAG_PAL || desc->flags & AV_PIX_FMT_FLAG_PSEUDOPAL) + if (desc->flags & AV_PIX_FMT_FLAG_PAL || desc->flags & AV_PIX_FMT_FLAG_PSEUDOPAL) { avpriv_set_systematic_pal2((uint32_t*)pointers[1], pix_fmt); + if (align < 4) { + av_log(NULL, AV_LOG_ERROR, "Formats with a palette require a minimum alignment of 4\n"); + return AVERROR(EINVAL); + } + } + + if ((desc->flags & AV_PIX_FMT_FLAG_PAL || + desc->flags & AV_PIX_FMT_FLAG_PSEUDOPAL) && + pointers[1] - pointers[0] > linesizes[0] * h) { + /* zero-initialize the padding before the palette */ + memset(pointers[0] + linesizes[0] * h, 0, + pointers[1] - pointers[0] - linesizes[0] * h); + } return ret; } @@ -218,11 +240,21 @@ typedef struct ImgUtils { void *log_ctx; } ImgUtils; -static const AVClass imgutils_class = { "IMGUTILS", av_default_item_name, NULL, LIBAVUTIL_VERSION_INT, offsetof(ImgUtils, log_offset), offsetof(ImgUtils, log_ctx) }; +static const AVClass imgutils_class = { + .class_name = "IMGUTILS", + .item_name = av_default_item_name, + .version = LIBAVUTIL_VERSION_INT, + .log_level_offset_offset = offsetof(ImgUtils, log_offset), + .parent_log_context_offset = offsetof(ImgUtils, log_ctx), +}; int av_image_check_size(unsigned int w, unsigned int h, int log_offset, void *log_ctx) { - ImgUtils imgutils = { &imgutils_class, log_offset, log_ctx }; + ImgUtils imgutils = { + .class = &imgutils_class, + .log_offset = log_offset, + .log_ctx = log_ctx, + }; if ((int)w>0 && (int)h>0 && (w+128)*(uint64_t)(h+128) < INT_MAX/8) return 0; @@ -235,7 +267,7 @@ int av_image_check_sar(unsigned int w, unsigned int h, AVRational sar) { int64_t scaled_dim; - if (!sar.den) + if (sar.den <= 0 || sar.num < 0) return AVERROR(EINVAL); if (!sar.num || sar.num == sar.den) @@ -258,6 +290,8 @@ void av_image_copy_plane(uint8_t *dst, int dst_linesize, { if (!dst || !src) return; + av_assert0(abs(src_linesize) >= bytewidth); + av_assert0(abs(dst_linesize) >= bytewidth); for (;height > 0; height--) { memcpy(dst, src, bytewidth); dst += dst_linesize; @@ -290,6 +324,10 @@ void av_image_copy(uint8_t *dst_data[4], int dst_linesizes[4], for (i = 0; i < planes_nb; i++) { int h = height; int bwidth = av_image_get_linesize(pix_fmt, width, i); + if (bwidth < 0) { + av_log(NULL, AV_LOG_ERROR, "av_image_get_linesize failed\n"); + return; + } if (i == 1 || i == 2) { h = AV_CEIL_RSHIFT(height, desc->log2_chroma_h); } @@ -317,7 +355,7 @@ int av_image_fill_arrays(uint8_t *dst_data[4], int dst_linesize[4], for (i = 0; i < 4; i++) dst_linesize[i] = FFALIGN(dst_linesize[i], align); - return av_image_fill_pointers(dst_data, pix_fmt, height, src, dst_linesize); + return av_image_fill_pointers(dst_data, pix_fmt, height, (uint8_t *)src, dst_linesize); } int av_image_get_buffer_size(enum AVPixelFormat pix_fmt, @@ -328,7 +366,7 @@ int av_image_get_buffer_size(enum AVPixelFormat pix_fmt, int ret; const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt); if (!desc) - return AVERROR_BUG; + return AVERROR(EINVAL); ret = av_image_check_size(width, height, 0, NULL); if (ret < 0) @@ -336,7 +374,7 @@ int av_image_get_buffer_size(enum AVPixelFormat pix_fmt, // do not include palette for these pseudo-paletted formats if (desc->flags & AV_PIX_FMT_FLAG_PSEUDOPAL) - return width * height; + return FFALIGN(width, align) * height; return av_image_fill_arrays(data, linesize, NULL, pix_fmt, width, height, align); @@ -351,6 +389,7 @@ int av_image_copy_to_buffer(uint8_t *dst, int dst_size, int i, j, nb_planes = 0, linesize[4]; int size = av_image_get_buffer_size(pix_fmt, width, height, align); const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt); + int ret; if (size > dst_size || size < 0 || !desc) return AVERROR(EINVAL); @@ -360,7 +399,9 @@ int av_image_copy_to_buffer(uint8_t *dst, int dst_size, nb_planes++; - av_image_fill_linesizes(linesize, pix_fmt, width); + ret = av_image_fill_linesizes(linesize, pix_fmt, width); + av_assert0(ret >= 0); // was checked previously + for (i = 0; i < nb_planes; i++) { int h, shift = (i == 1 || i == 2) ? desc->log2_chroma_h : 0; const uint8_t *src = src_data[i]; @@ -373,9 +414,12 @@ int av_image_copy_to_buffer(uint8_t *dst, int dst_size, } } - if (desc->flags & AV_PIX_FMT_FLAG_PAL) - memcpy((unsigned char *)(((size_t)dst + 3) & ~3), - src_data[1], 256 * 4); + if (desc->flags & AV_PIX_FMT_FLAG_PAL) { + uint32_t *d32 = (uint32_t *)dst; + + for (i = 0; i<256; i++) + AV_WL32(d32 + i, AV_RN32(src_data[1] + 4*i)); + } return size; } diff --git a/libavutil/imgutils.h b/libavutil/imgutils.h index 04d6138f67..23282a38fa 100644 --- a/libavutil/imgutils.h +++ b/libavutil/imgutils.h @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -100,6 +100,9 @@ int av_image_alloc(uint8_t *pointers[4], int linesizes[4], * The first byte of each successive line is separated by *_linesize * bytes. * + * bytewidth must be contained by both absolute values of dst_linesize + * and src_linesize, otherwise the function behavior is undefined. + * * @param dst_linesize linesize for the image plane in dst * @param src_linesize linesize for the image plane in src */ diff --git a/libavutil/integer.c b/libavutil/integer.c new file mode 100644 index 0000000000..6d6855fa1b --- /dev/null +++ b/libavutil/integer.c @@ -0,0 +1,204 @@ +/* + * arbitrary precision integers + * Copyright (c) 2004 Michael Niedermayer <michaelni@gmx.at> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +/** + * @file + * arbitrary precision integers + * @author Michael Niedermayer <michaelni@gmx.at> + */ + +#include "common.h" +#include "integer.h" +#include "avassert.h" + +static const AVInteger zero_i; + +AVInteger av_add_i(AVInteger a, AVInteger b){ + int i, carry=0; + + for(i=0; i<AV_INTEGER_SIZE; i++){ + carry= (carry>>16) + a.v[i] + b.v[i]; + a.v[i]= carry; + } + return a; +} + +AVInteger av_sub_i(AVInteger a, AVInteger b){ + int i, carry=0; + + for(i=0; i<AV_INTEGER_SIZE; i++){ + carry= (carry>>16) + a.v[i] - b.v[i]; + a.v[i]= carry; + } + return a; +} + +int av_log2_i(AVInteger a){ + int i; + + for(i=AV_INTEGER_SIZE-1; i>=0; i--){ + if(a.v[i]) + return av_log2_16bit(a.v[i]) + 16*i; + } + return -1; +} + +AVInteger av_mul_i(AVInteger a, AVInteger b){ + AVInteger out; + int i, j; + int na= (av_log2_i(a)+16) >> 4; + int nb= (av_log2_i(b)+16) >> 4; + + memset(&out, 0, sizeof(out)); + + for(i=0; i<na; i++){ + unsigned int carry=0; + + if(a.v[i]) + for(j=i; j<AV_INTEGER_SIZE && j-i<=nb; j++){ + carry= (carry>>16) + out.v[j] + a.v[i]*b.v[j-i]; + out.v[j]= carry; + } + } + + return out; +} + +int av_cmp_i(AVInteger a, AVInteger b){ + int i; + int v= (int16_t)a.v[AV_INTEGER_SIZE-1] - (int16_t)b.v[AV_INTEGER_SIZE-1]; + if(v) return (v>>16)|1; + + for(i=AV_INTEGER_SIZE-2; i>=0; i--){ + int v= a.v[i] - b.v[i]; + if(v) return (v>>16)|1; + } + return 0; +} + +AVInteger av_shr_i(AVInteger a, int s){ + AVInteger out; + int i; + + for(i=0; i<AV_INTEGER_SIZE; i++){ + unsigned int index= i + (s>>4); + unsigned int v=0; + if(index+1<AV_INTEGER_SIZE) v = a.v[index+1]<<16; + if(index <AV_INTEGER_SIZE) v+= a.v[index ]; + out.v[i]= v >> (s&15); + } + return out; +} + +AVInteger av_mod_i(AVInteger *quot, AVInteger a, AVInteger b){ + int i= av_log2_i(a) - av_log2_i(b); + AVInteger quot_temp; + if(!quot) quot = "_temp; + + if ((int16_t)a.v[AV_INTEGER_SIZE-1] < 0) { + a = av_mod_i(quot, av_sub_i(zero_i, a), b); + *quot = av_sub_i(zero_i, *quot); + return av_sub_i(zero_i, a); + } + + av_assert2((int16_t)a.v[AV_INTEGER_SIZE-1] >= 0 && (int16_t)b.v[AV_INTEGER_SIZE-1] >= 0); + av_assert2(av_log2_i(b)>=0); + + if(i > 0) + b= av_shr_i(b, -i); + + memset(quot, 0, sizeof(AVInteger)); + + while(i-- >= 0){ + *quot= av_shr_i(*quot, -1); + if(av_cmp_i(a, b) >= 0){ + a= av_sub_i(a, b); + quot->v[0] += 1; + } + b= av_shr_i(b, 1); + } + return a; +} + +AVInteger av_div_i(AVInteger a, AVInteger b){ + AVInteger quot; + av_mod_i(", a, b); + return quot; +} + +AVInteger av_int2i(int64_t a){ + AVInteger out; + int i; + + for(i=0; i<AV_INTEGER_SIZE; i++){ + out.v[i]= a; + a>>=16; + } + return out; +} + +int64_t av_i2int(AVInteger a){ + int i; + int64_t out=(int8_t)a.v[AV_INTEGER_SIZE-1]; + + for(i= AV_INTEGER_SIZE-2; i>=0; i--){ + out = (out<<16) + a.v[i]; + } + return out; +} + +#ifdef TEST + +const uint8_t ff_log2_tab[256]={ + 0,0,1,1,2,2,2,2,3,3,3,3,3,3,3,3,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4, + 5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5, + 6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6, + 6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6, + 7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7, + 7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7, + 7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7, + 7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7 +}; + +int main(void){ + int64_t a,b; + + for(a=7; a<256*256*256; a+=13215){ + for(b=3; b<256*256*256; b+=27118){ + AVInteger ai= av_int2i(a); + AVInteger bi= av_int2i(b); + + av_assert0(av_i2int(ai) == a); + av_assert0(av_i2int(bi) == b); + av_assert0(av_i2int(av_add_i(ai,bi)) == a+b); + av_assert0(av_i2int(av_sub_i(ai,bi)) == a-b); + av_assert0(av_i2int(av_mul_i(ai,bi)) == a*b); + av_assert0(av_i2int(av_shr_i(ai, 9)) == a>>9); + av_assert0(av_i2int(av_shr_i(ai,-9)) == a<<9); + av_assert0(av_i2int(av_shr_i(ai, 17)) == a>>17); + av_assert0(av_i2int(av_shr_i(ai,-17)) == a<<17); + av_assert0(av_log2_i(ai) == av_log2(a)); + av_assert0(av_i2int(av_div_i(ai,bi)) == a/b); + } + } + return 0; +} +#endif diff --git a/libavutil/integer.h b/libavutil/integer.h new file mode 100644 index 0000000000..45f733c04c --- /dev/null +++ b/libavutil/integer.h @@ -0,0 +1,86 @@ +/* + * arbitrary precision integers + * Copyright (c) 2004 Michael Niedermayer <michaelni@gmx.at> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +/** + * @file + * arbitrary precision integers + * @author Michael Niedermayer <michaelni@gmx.at> + */ + +#ifndef AVUTIL_INTEGER_H +#define AVUTIL_INTEGER_H + +#include <stdint.h> +#include "common.h" + +#define AV_INTEGER_SIZE 8 + +typedef struct AVInteger{ + uint16_t v[AV_INTEGER_SIZE]; +} AVInteger; + +AVInteger av_add_i(AVInteger a, AVInteger b) av_const; +AVInteger av_sub_i(AVInteger a, AVInteger b) av_const; + +/** + * Return the rounded-down value of the base 2 logarithm of the given + * AVInteger. This is simply the index of the most significant bit + * which is 1, or 0 if all bits are 0. + */ +int av_log2_i(AVInteger a) av_const; +AVInteger av_mul_i(AVInteger a, AVInteger b) av_const; + +/** + * Return 0 if a==b, 1 if a>b and -1 if a<b. + */ +int av_cmp_i(AVInteger a, AVInteger b) av_const; + +/** + * bitwise shift + * @param s the number of bits by which the value should be shifted right, + may be negative for shifting left + */ +AVInteger av_shr_i(AVInteger a, int s) av_const; + +/** + * Return a % b. + * @param quot a/b will be stored here. + */ +AVInteger av_mod_i(AVInteger *quot, AVInteger a, AVInteger b); + +/** + * Return a/b. + */ +AVInteger av_div_i(AVInteger a, AVInteger b) av_const; + +/** + * Convert the given int64_t to an AVInteger. + */ +AVInteger av_int2i(int64_t a) av_const; + +/** + * Convert the given AVInteger to an int64_t. + * If the AVInteger is too large to fit into an int64_t, + * then only the least significant 64 bits will be used. + */ +int64_t av_i2int(AVInteger a) av_const; + +#endif /* AVUTIL_INTEGER_H */ diff --git a/libavutil/internal.h b/libavutil/internal.h index b9be333b47..61784b5128 100644 --- a/libavutil/internal.h +++ b/libavutil/internal.h @@ -1,20 +1,20 @@ /* * copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -36,16 +36,19 @@ #include <assert.h> #include "config.h" #include "attributes.h" +#include "timer.h" +#include "cpu.h" #include "dict.h" #include "macros.h" #include "pixfmt.h" +#include "version.h" #if ARCH_X86 # include "x86/emms.h" #endif #ifndef emms_c -# define emms_c() +# define emms_c() while(0) #endif #ifndef attribute_align_arg @@ -63,7 +66,7 @@ #endif #if HAVE_PRAGMA_DEPRECATED -# if defined(__ICL) +# if defined(__ICL) || defined (__INTEL_COMPILER) # define FF_DISABLE_DEPRECATION_WARNINGS __pragma(warning(push)) __pragma(warning(disable:1478)) # define FF_ENABLE_DEPRECATION_WARNINGS __pragma(warning(pop)) # elif defined(_MSC_VER) @@ -78,6 +81,13 @@ # define FF_ENABLE_DEPRECATION_WARNINGS #endif + +#define FF_MEMORY_POISON 0x2a + +#define MAKE_ACCESSORS(str, name, type, field) \ + type av_##name##_get_##field(const str *s) { return s->field; } \ + void av_##name##_set_##field(str *s, type v) { s->field = v; } + // Some broken preprocessors need a second expansion // to be forced to tokenize __VA_ARGS__ #define E1(x) x @@ -111,6 +121,12 @@ # define LOCAL_ALIGNED_16(t, v, ...) LOCAL_ALIGNED(16, t, v, __VA_ARGS__) #endif +#if HAVE_LOCAL_ALIGNED_32 +# define LOCAL_ALIGNED_32(t, v, ...) E1(LOCAL_ALIGNED_D(32, t, v, __VA_ARGS__,,)) +#else +# define LOCAL_ALIGNED_32(t, v, ...) LOCAL_ALIGNED(32, t, v, __VA_ARGS__) +#endif + #define FF_ALLOC_OR_GOTO(ctx, p, size, label)\ {\ p = av_malloc(size);\ @@ -129,6 +145,24 @@ }\ } +#define FF_ALLOC_ARRAY_OR_GOTO(ctx, p, nelem, elsize, label)\ +{\ + p = av_malloc_array(nelem, elsize);\ + if (!p) {\ + av_log(ctx, AV_LOG_ERROR, "Cannot allocate memory.\n");\ + goto label;\ + }\ +} + +#define FF_ALLOCZ_ARRAY_OR_GOTO(ctx, p, nelem, elsize, label)\ +{\ + p = av_mallocz_array(nelem, elsize);\ + if (!p) {\ + av_log(ctx, AV_LOG_ERROR, "Cannot allocate memory.\n");\ + goto label;\ + }\ +} + #include "libm.h" /** @@ -142,12 +176,11 @@ # define NULL_IF_CONFIG_SMALL(x) x #endif - /** * Define a function with only the non-default version specified. * * On systems with ELF shared libraries, all symbols exported from - * Libav libraries are tagged with the name and major version of the + * FFmpeg libraries are tagged with the name and major version of the * library to which they belong. If a function is moved from one * library to another, a wrapper must be retained in the original * location to preserve binary compatibility. @@ -206,18 +239,97 @@ void avpriv_request_sample(void *avc, #if HAVE_LIBC_MSVCRT #include <crtversion.h> #if defined(_VC_CRT_MAJOR_VERSION) && _VC_CRT_MAJOR_VERSION < 14 -#pragma comment(linker, "/include:"EXTERN_PREFIX"avpriv_strtod") -#pragma comment(linker, "/include:"EXTERN_PREFIX"avpriv_snprintf") +#pragma comment(linker, "/include:" EXTERN_PREFIX "avpriv_strtod") +#pragma comment(linker, "/include:" EXTERN_PREFIX "avpriv_snprintf") #endif #define avpriv_open ff_open +#define avpriv_tempfile ff_tempfile +#define PTRDIFF_SPECIFIER "Id" +#define SIZE_SPECIFIER "Iu" +#else +#define PTRDIFF_SPECIFIER "td" +#define SIZE_SPECIFIER "zu" +#endif + +#ifdef DEBUG +# define ff_dlog(ctx, ...) av_log(ctx, AV_LOG_DEBUG, __VA_ARGS__) +#else +# define ff_dlog(ctx, ...) do { if (0) av_log(ctx, AV_LOG_DEBUG, __VA_ARGS__); } while (0) +#endif + +/** + * Clip and convert a double value into the long long amin-amax range. + * This function is needed because conversion of floating point to integers when + * it does not fit in the integer's representation does not necessarily saturate + * correctly (usually converted to a cvttsd2si on x86) which saturates numbers + * > INT64_MAX to INT64_MIN. The standard marks such conversions as undefined + * behavior, allowing this sort of mathematically bogus conversions. This provides + * a safe alternative that is slower obviously but assures safety and better + * mathematical behavior. + * @param a value to clip + * @param amin minimum value of the clip range + * @param amax maximum value of the clip range + * @return clipped value + */ +static av_always_inline av_const int64_t ff_rint64_clip(double a, int64_t amin, int64_t amax) +{ + int64_t res; +#if defined(HAVE_AV_CONFIG_H) && defined(ASSERT_LEVEL) && ASSERT_LEVEL >= 2 + if (amin > amax) abort(); #endif + // INT64_MAX+1,INT64_MIN are exactly representable as IEEE doubles + // do range checks first + if (a >= 9223372036854775808.0) + return amax; + if (a <= -9223372036854775808.0) + return amin; + + // safe to call llrint and clip accordingly + res = llrint(a); + if (res > amax) + return amax; + if (res < amin) + return amin; + return res; +} /** * A wrapper for open() setting O_CLOEXEC. */ +av_warn_unused_result int avpriv_open(const char *filename, int flags, ...); +/** + * Wrapper to work around the lack of mkstemp() on mingw. + * Also, tries to create file in /tmp first, if possible. + * *prefix can be a character constant; *filename will be allocated internally. + * @return file descriptor of opened file (or negative value corresponding to an + * AVERROR code on error) + * and opened file name in **filename. + * @note On very old libcs it is necessary to set a secure umask before + * calling this, av_tempfile() can't call umask itself as it is used in + * libraries and could interfere with the calling application. + */ +int avpriv_tempfile(const char *prefix, char **filename, int log_offset, void *log_ctx); + int avpriv_set_systematic_pal2(uint32_t pal[256], enum AVPixelFormat pix_fmt); +static av_always_inline av_const int avpriv_mirror(int x, int w) +{ + if (!w) + return 0; + + while ((unsigned)x > (unsigned)w) { + x = -x; + if (x < 0) + x += 2 * w; + } + return x; +} + +void ff_check_pixfmt_descriptors(void); + +extern const uint8_t ff_reverse[256]; + #endif /* AVUTIL_INTERNAL_H */ diff --git a/libavutil/intfloat.h b/libavutil/intfloat.h index 38d26ad87e..fe3d7ec4a5 100644 --- a/libavutil/intfloat.h +++ b/libavutil/intfloat.h @@ -1,20 +1,20 @@ /* * Copyright (c) 2011 Mans Rullgard * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/intmath.c b/libavutil/intmath.c index 8db425c6e9..b0c00e1cad 100644 --- a/libavutil/intmath.c +++ b/libavutil/intmath.c @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -32,8 +32,3 @@ int av_log2_16bit(unsigned v) { return ff_log2_16bit(v); } - -int av_ctz(int v) -{ - return ff_ctz(v); -} diff --git a/libavutil/intmath.h b/libavutil/intmath.h index a5ee6525ee..9573109e9d 100644 --- a/libavutil/intmath.h +++ b/libavutil/intmath.h @@ -1,20 +1,20 @@ /* * Copyright (c) 2010 Mans Rullgard <mans@mansr.com> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -29,22 +29,20 @@ #if ARCH_ARM # include "arm/intmath.h" #endif +#if ARCH_X86 +# include "x86/intmath.h" +#endif -/** - * @addtogroup lavu_internal - * @{ - */ - -#if HAVE_FAST_CLZ && AV_GCC_VERSION_AT_LEAST(3,4) - +#if HAVE_FAST_CLZ +#if AV_GCC_VERSION_AT_LEAST(3,4) #ifndef ff_log2 # define ff_log2(x) (31 - __builtin_clz((x)|1)) # ifndef ff_log2_16bit # define ff_log2_16bit av_log2 # endif #endif /* ff_log2 */ - #endif /* AV_GCC_VERSION_AT_LEAST(3,4) */ +#endif extern const uint8_t ff_log2_tab[256]; @@ -86,59 +84,80 @@ static av_always_inline av_const int ff_log2_16bit_c(unsigned int v) #define av_log2_16bit ff_log2_16bit /** - * @} - */ - -/** * @addtogroup lavu_math * @{ */ -#if HAVE_FAST_CLZ && AV_GCC_VERSION_AT_LEAST(3,4) +#if HAVE_FAST_CLZ +#if AV_GCC_VERSION_AT_LEAST(3,4) #ifndef ff_ctz #define ff_ctz(v) __builtin_ctz(v) #endif +#ifndef ff_ctzll +#define ff_ctzll(v) __builtin_ctzll(v) +#endif +#ifndef ff_clz +#define ff_clz(v) __builtin_clz(v) +#endif +#endif #endif #ifndef ff_ctz #define ff_ctz ff_ctz_c +/** + * Trailing zero bit count. + * + * @param v input value. If v is 0, the result is undefined. + * @return the number of trailing 0-bits + */ +/* We use the De-Bruijn method outlined in: + * http://supertech.csail.mit.edu/papers/debruijn.pdf. */ static av_always_inline av_const int ff_ctz_c(int v) { - int c; + static const uint8_t debruijn_ctz32[32] = { + 0, 1, 28, 2, 29, 14, 24, 3, 30, 22, 20, 15, 25, 17, 4, 8, + 31, 27, 13, 23, 21, 19, 16, 7, 26, 12, 18, 6, 11, 5, 10, 9 + }; + return debruijn_ctz32[(uint32_t)((v & -v) * 0x077CB531U) >> 27]; +} +#endif - if (v & 0x1) - return 0; +#ifndef ff_ctzll +#define ff_ctzll ff_ctzll_c +/* We use the De-Bruijn method outlined in: + * http://supertech.csail.mit.edu/papers/debruijn.pdf. */ +static av_always_inline av_const int ff_ctzll_c(long long v) +{ + static const uint8_t debruijn_ctz64[64] = { + 0, 1, 2, 53, 3, 7, 54, 27, 4, 38, 41, 8, 34, 55, 48, 28, + 62, 5, 39, 46, 44, 42, 22, 9, 24, 35, 59, 56, 49, 18, 29, 11, + 63, 52, 6, 26, 37, 40, 33, 47, 61, 45, 43, 21, 23, 58, 17, 10, + 51, 25, 36, 32, 60, 20, 57, 16, 50, 31, 19, 15, 30, 14, 13, 12 + }; + return debruijn_ctz64[(uint64_t)((v & -v) * 0x022FDD63CC95386DU) >> 58]; +} +#endif - c = 1; - if (!(v & 0xffff)) { - v >>= 16; - c += 16; - } - if (!(v & 0xff)) { - v >>= 8; - c += 8; - } - if (!(v & 0xf)) { - v >>= 4; - c += 4; - } - if (!(v & 0x3)) { - v >>= 2; - c += 2; +#ifndef ff_clz +#define ff_clz ff_clz_c +static av_always_inline av_const unsigned ff_clz_c(unsigned x) +{ + unsigned i = sizeof(x) * 8; + + while (x) { + x >>= 1; + i--; } - c -= v & 0x1; - return c; + return i; } #endif -/** - * Trailing zero bit count. - * - * @param v input value. If v is 0, the result is undefined. - * @return the number of trailing 0-bits - */ -int av_ctz(int v); +#if AV_GCC_VERSION_AT_LEAST(3,4) +#ifndef av_parity +#define av_parity __builtin_parity +#endif +#endif /** * @} diff --git a/libavutil/intreadwrite.h b/libavutil/intreadwrite.h index f77fd60f38..51fbe30a23 100644 --- a/libavutil/intreadwrite.h +++ b/libavutil/intreadwrite.h @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -47,9 +47,17 @@ typedef union { /* * Arch-specific headers can provide any combination of - * AV_[RW][BLN](16|24|32|64) and AV_(COPY|SWAP|ZERO)(64|128) macros. + * AV_[RW][BLN](16|24|32|48|64) and AV_(COPY|SWAP|ZERO)(64|128) macros. * Preprocessor symbols must be defined, even if these are implemented * as inline functions. + * + * R/W means read/write, B/L/N means big/little/native endianness. + * The following macros require aligned access, compared to their + * unaligned variants: AV_(COPY|SWAP|ZERO)(64|128), AV_[RW]N[8-64]A. + * Incorrect usage may range from abysmal performance to crash + * depending on the platform. + * + * The unaligned variants are AV_[RW][BLN][8-64] and AV_COPY*U. */ #ifdef HAVE_AV_CONFIG_H @@ -114,6 +122,18 @@ typedef union { # define AV_WN32(p, v) AV_WB32(p, v) # endif +# if defined(AV_RN48) && !defined(AV_RB48) +# define AV_RB48(p) AV_RN48(p) +# elif !defined(AV_RN48) && defined(AV_RB48) +# define AV_RN48(p) AV_RB48(p) +# endif + +# if defined(AV_WN48) && !defined(AV_WB48) +# define AV_WB48(p, v) AV_WN48(p, v) +# elif !defined(AV_WN48) && defined(AV_WB48) +# define AV_WN48(p, v) AV_WB48(p, v) +# endif + # if defined(AV_RN64) && !defined(AV_RB64) # define AV_RB64(p) AV_RN64(p) # elif !defined(AV_RN64) && defined(AV_RB64) @@ -164,6 +184,18 @@ typedef union { # define AV_WN32(p, v) AV_WL32(p, v) # endif +# if defined(AV_RN48) && !defined(AV_RL48) +# define AV_RL48(p) AV_RN48(p) +# elif !defined(AV_RN48) && defined(AV_RL48) +# define AV_RN48(p) AV_RL48(p) +# endif + +# if defined(AV_WN48) && !defined(AV_WL48) +# define AV_WL48(p, v) AV_WN48(p, v) +# elif !defined(AV_WN48) && defined(AV_WL48) +# define AV_WN48(p, v) AV_WL48(p, v) +# endif + # if defined(AV_RN64) && !defined(AV_RL64) # define AV_RL64(p) AV_RN64(p) # elif !defined(AV_RN64) && defined(AV_RL64) @@ -210,7 +242,8 @@ union unaligned_16 { uint16_t l; } __attribute__((packed)) av_alias; ((const uint8_t*)(x))[1]) #endif #ifndef AV_WB16 -# define AV_WB16(p, d) do { \ +# define AV_WB16(p, darg) do { \ + unsigned d = (darg); \ ((uint8_t*)(p))[1] = (d); \ ((uint8_t*)(p))[0] = (d)>>8; \ } while(0) @@ -222,7 +255,8 @@ union unaligned_16 { uint16_t l; } __attribute__((packed)) av_alias; ((const uint8_t*)(x))[0]) #endif #ifndef AV_WL16 -# define AV_WL16(p, d) do { \ +# define AV_WL16(p, darg) do { \ + unsigned d = (darg); \ ((uint8_t*)(p))[0] = (d); \ ((uint8_t*)(p))[1] = (d)>>8; \ } while(0) @@ -236,7 +270,8 @@ union unaligned_16 { uint16_t l; } __attribute__((packed)) av_alias; ((const uint8_t*)(x))[3]) #endif #ifndef AV_WB32 -# define AV_WB32(p, d) do { \ +# define AV_WB32(p, darg) do { \ + unsigned d = (darg); \ ((uint8_t*)(p))[3] = (d); \ ((uint8_t*)(p))[2] = (d)>>8; \ ((uint8_t*)(p))[1] = (d)>>16; \ @@ -252,7 +287,8 @@ union unaligned_16 { uint16_t l; } __attribute__((packed)) av_alias; ((const uint8_t*)(x))[0]) #endif #ifndef AV_WL32 -# define AV_WL32(p, d) do { \ +# define AV_WL32(p, darg) do { \ + unsigned d = (darg); \ ((uint8_t*)(p))[0] = (d); \ ((uint8_t*)(p))[1] = (d)>>8; \ ((uint8_t*)(p))[2] = (d)>>16; \ @@ -272,7 +308,8 @@ union unaligned_16 { uint16_t l; } __attribute__((packed)) av_alias; (uint64_t)((const uint8_t*)(x))[7]) #endif #ifndef AV_WB64 -# define AV_WB64(p, d) do { \ +# define AV_WB64(p, darg) do { \ + uint64_t d = (darg); \ ((uint8_t*)(p))[7] = (d); \ ((uint8_t*)(p))[6] = (d)>>8; \ ((uint8_t*)(p))[5] = (d)>>16; \ @@ -296,7 +333,8 @@ union unaligned_16 { uint16_t l; } __attribute__((packed)) av_alias; (uint64_t)((const uint8_t*)(x))[0]) #endif #ifndef AV_WL64 -# define AV_WL64(p, d) do { \ +# define AV_WL64(p, darg) do { \ + uint64_t d = (darg); \ ((uint8_t*)(p))[0] = (d); \ ((uint8_t*)(p))[1] = (d)>>8; \ ((uint8_t*)(p))[2] = (d)>>16; \ @@ -430,6 +468,48 @@ union unaligned_16 { uint16_t l; } __attribute__((packed)) av_alias; } while(0) #endif +#ifndef AV_RB48 +# define AV_RB48(x) \ + (((uint64_t)((const uint8_t*)(x))[0] << 40) | \ + ((uint64_t)((const uint8_t*)(x))[1] << 32) | \ + ((uint64_t)((const uint8_t*)(x))[2] << 24) | \ + ((uint64_t)((const uint8_t*)(x))[3] << 16) | \ + ((uint64_t)((const uint8_t*)(x))[4] << 8) | \ + (uint64_t)((const uint8_t*)(x))[5]) +#endif +#ifndef AV_WB48 +# define AV_WB48(p, darg) do { \ + uint64_t d = (darg); \ + ((uint8_t*)(p))[5] = (d); \ + ((uint8_t*)(p))[4] = (d)>>8; \ + ((uint8_t*)(p))[3] = (d)>>16; \ + ((uint8_t*)(p))[2] = (d)>>24; \ + ((uint8_t*)(p))[1] = (d)>>32; \ + ((uint8_t*)(p))[0] = (d)>>40; \ + } while(0) +#endif + +#ifndef AV_RL48 +# define AV_RL48(x) \ + (((uint64_t)((const uint8_t*)(x))[5] << 40) | \ + ((uint64_t)((const uint8_t*)(x))[4] << 32) | \ + ((uint64_t)((const uint8_t*)(x))[3] << 24) | \ + ((uint64_t)((const uint8_t*)(x))[2] << 16) | \ + ((uint64_t)((const uint8_t*)(x))[1] << 8) | \ + (uint64_t)((const uint8_t*)(x))[0]) +#endif +#ifndef AV_WL48 +# define AV_WL48(p, darg) do { \ + uint64_t d = (darg); \ + ((uint8_t*)(p))[0] = (d); \ + ((uint8_t*)(p))[1] = (d)>>8; \ + ((uint8_t*)(p))[2] = (d)>>16; \ + ((uint8_t*)(p))[3] = (d)>>24; \ + ((uint8_t*)(p))[4] = (d)>>32; \ + ((uint8_t*)(p))[5] = (d)>>40; \ + } while(0) +#endif + /* * The AV_[RW]NA macros access naturally aligned data * in a type-safe way. diff --git a/libavutil/lfg-test.c b/libavutil/lfg-test.c index 92b4259a3c..2b6f3828dc 100644 --- a/libavutil/lfg-test.c +++ b/libavutil/lfg-test.c @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -29,8 +29,10 @@ int main(void) av_lfg_init(&state, 0xdeadbeef); for (j = 0; j < 10000; j++) { START_TIMER - for (i = 0; i < 624; i++) + for (i = 0; i < 624; i++) { + //av_log(NULL, AV_LOG_ERROR, "%X\n", av_lfg_get(&state)); x += av_lfg_get(&state); + } STOP_TIMER("624 calls of av_lfg_get"); } av_log(NULL, AV_LOG_ERROR, "final value:%X\n", x); @@ -39,17 +41,33 @@ int main(void) { double mean = 1000; double stddev = 53; + double samp_mean = 0.0, samp_stddev = 0.0; + double samp0, samp1; av_lfg_init(&state, 42); for (i = 0; i < 1000; i += 2) { double bmg_out[2]; av_bmg_get(&state, bmg_out); + samp0 = bmg_out[0] * stddev + mean; + samp1 = bmg_out[1] * stddev + mean; + samp_mean += samp0 + samp1; + samp_stddev += samp0 * samp0 + samp1 * samp1; av_log(NULL, AV_LOG_INFO, "%f\n%f\n", - bmg_out[0] * stddev + mean, - bmg_out[1] * stddev + mean); + samp0, + samp1); } + /* TODO: add proper normality test */ + samp_mean /= 1000; + samp_stddev /= 999; + samp_stddev -= (1000.0/999.0)*samp_mean*samp_mean; + samp_stddev = sqrt(samp_stddev); + av_log(NULL, AV_LOG_INFO, "sample mean : %f\n" + "true mean : %f\n" + "sample stddev: %f\n" + "true stddev : %f\n", + samp_mean, mean, samp_stddev, stddev); } return 0; diff --git a/libavutil/lfg.c b/libavutil/lfg.c index 20900570a2..08a4f678e2 100644 --- a/libavutil/lfg.c +++ b/libavutil/lfg.c @@ -2,20 +2,20 @@ * Lagged Fibonacci PRNG * Copyright (c) 2008 Michael Niedermayer * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/lfg.h b/libavutil/lfg.h index 5e526c1dae..ec90562cf2 100644 --- a/libavutil/lfg.h +++ b/libavutil/lfg.h @@ -2,20 +2,20 @@ * Lagged Fibonacci PRNG * Copyright (c) 2008 Michael Niedermayer * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/libm.h b/libavutil/libm.h index d6c2cf8623..a819962391 100644 --- a/libavutil/libm.h +++ b/libavutil/libm.h @@ -1,18 +1,19 @@ /* - * This file is part of Libav. + * erf function: Copyright (c) 2006 John Maddock + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -28,42 +29,263 @@ #include "config.h" #include "attributes.h" #include "intfloat.h" +#include "mathematics.h" + +#if HAVE_MIPSFPU && HAVE_INLINE_ASM +#include "libavutil/mips/libm_mips.h" +#endif /* HAVE_MIPSFPU && HAVE_INLINE_ASM*/ #if !HAVE_ATANF #undef atanf #define atanf(x) ((float)atan(x)) -#endif +#endif /* HAVE_ATANF */ #if !HAVE_ATAN2F #undef atan2f #define atan2f(y, x) ((float)atan2(y, x)) -#endif +#endif /* HAVE_ATAN2F */ #if !HAVE_POWF #undef powf #define powf(x, y) ((float)pow(x, y)) -#endif +#endif /* HAVE_POWF */ + +#if !HAVE_CBRT +static av_always_inline double cbrt(double x) +{ + return x < 0 ? -pow(-x, 1.0 / 3.0) : pow(x, 1.0 / 3.0); +} +#endif /* HAVE_CBRT */ #if !HAVE_CBRTF static av_always_inline float cbrtf(float x) { return x < 0 ? -powf(-x, 1.0 / 3.0) : powf(x, 1.0 / 3.0); } -#endif +#endif /* HAVE_CBRTF */ + +#if !HAVE_COPYSIGN +static av_always_inline double copysign(double x, double y) +{ + uint64_t vx = av_double2int(x); + uint64_t vy = av_double2int(y); + return av_int2double((vx & UINT64_C(0x7fffffffffffffff)) | (vy & UINT64_C(0x8000000000000000))); +} +#endif /* HAVE_COPYSIGN */ #if !HAVE_COSF #undef cosf #define cosf(x) ((float)cos(x)) +#endif /* HAVE_COSF */ + +#if !HAVE_ERF +static inline double ff_eval_poly(const double *coeff, int size, double x) { + double sum = coeff[size-1]; + int i; + for (i = size-2; i >= 0; --i) { + sum *= x; + sum += coeff[i]; + } + return sum; +} + +/** + * erf function + * Algorithm taken from the Boost project, source: + * http://www.boost.org/doc/libs/1_46_1/boost/math/special_functions/erf.hpp + * Use, modification and distribution are subject to the + * Boost Software License, Version 1.0 (see notice below). + * Boost Software License - Version 1.0 - August 17th, 2003 +Permission is hereby granted, free of charge, to any person or organization +obtaining a copy of the software and accompanying documentation covered by +this license (the "Software") to use, reproduce, display, distribute, +execute, and transmit the Software, and to prepare derivative works of the +Software, and to permit third-parties to whom the Software is furnished to +do so, all subject to the following: + +The copyright notices in the Software and this entire statement, including +the above license grant, this restriction and the following disclaimer, +must be included in all copies of the Software, in whole or in part, and +all derivative works of the Software, unless such copies or derivative +works are solely in the form of machine-executable object code generated by +a source language processor. + +THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR +IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, +FITNESS FOR A PARTICULAR PURPOSE, TITLE AND NON-INFRINGEMENT. IN NO EVENT +SHALL THE COPYRIGHT HOLDERS OR ANYONE DISTRIBUTING THE SOFTWARE BE LIABLE +FOR ANY DAMAGES OR OTHER LIABILITY, WHETHER IN CONTRACT, TORT OR OTHERWISE, +ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER +DEALINGS IN THE SOFTWARE. + */ +static inline double erf(double z) +{ +#ifndef FF_ARRAY_ELEMS +#define FF_ARRAY_ELEMS(a) (sizeof(a) / sizeof((a)[0])) #endif + double result; + + /* handle the symmetry: erf(-x) = -erf(x) */ + if (z < 0) + return -erf(-z); + + /* branch based on range of z, and pick appropriate approximation */ + if (z == 0) + return 0; + else if (z < 1e-10) + return z * 1.125 + z * 0.003379167095512573896158903121545171688; + else if (z < 0.5) { + // Maximum Deviation Found: 1.561e-17 + // Expected Error Term: 1.561e-17 + // Maximum Relative Change in Control Points: 1.155e-04 + // Max Error found at double precision = 2.961182e-17 + + static const double y = 1.044948577880859375; + static const double p[] = { + 0.0834305892146531832907, + -0.338165134459360935041, + -0.0509990735146777432841, + -0.00772758345802133288487, + -0.000322780120964605683831, + }; + static const double q[] = { + 1, + 0.455004033050794024546, + 0.0875222600142252549554, + 0.00858571925074406212772, + 0.000370900071787748000569, + }; + double zz = z * z; + return z * (y + ff_eval_poly(p, FF_ARRAY_ELEMS(p), zz) / ff_eval_poly(q, FF_ARRAY_ELEMS(q), zz)); + } + /* here onwards compute erfc */ + else if (z < 1.5) { + // Maximum Deviation Found: 3.702e-17 + // Expected Error Term: 3.702e-17 + // Maximum Relative Change in Control Points: 2.845e-04 + // Max Error found at double precision = 4.841816e-17 + static const double y = 0.405935764312744140625; + static const double p[] = { + -0.098090592216281240205, + 0.178114665841120341155, + 0.191003695796775433986, + 0.0888900368967884466578, + 0.0195049001251218801359, + 0.00180424538297014223957, + }; + static const double q[] = { + 1, + 1.84759070983002217845, + 1.42628004845511324508, + 0.578052804889902404909, + 0.12385097467900864233, + 0.0113385233577001411017, + 0.337511472483094676155e-5, + }; + result = y + ff_eval_poly(p, FF_ARRAY_ELEMS(p), z - 0.5) / ff_eval_poly(q, FF_ARRAY_ELEMS(q), z - 0.5); + result *= exp(-z * z) / z; + return 1 - result; + } + else if (z < 2.5) { + // Max Error found at double precision = 6.599585e-18 + // Maximum Deviation Found: 3.909e-18 + // Expected Error Term: 3.909e-18 + // Maximum Relative Change in Control Points: 9.886e-05 + static const double y = 0.50672817230224609375; + static const double p[] = { + -0.0243500476207698441272, + 0.0386540375035707201728, + 0.04394818964209516296, + 0.0175679436311802092299, + 0.00323962406290842133584, + 0.000235839115596880717416, + }; + static const double q[] = { + 1, + 1.53991494948552447182, + 0.982403709157920235114, + 0.325732924782444448493, + 0.0563921837420478160373, + 0.00410369723978904575884, + }; + result = y + ff_eval_poly(p, FF_ARRAY_ELEMS(p), z - 1.5) / ff_eval_poly(q, FF_ARRAY_ELEMS(q), z - 1.5); + result *= exp(-z * z) / z; + return 1 - result; + } + else if (z < 4.5) { + // Maximum Deviation Found: 1.512e-17 + // Expected Error Term: 1.512e-17 + // Maximum Relative Change in Control Points: 2.222e-04 + // Max Error found at double precision = 2.062515e-17 + static const double y = 0.5405750274658203125; + static const double p[] = { + 0.00295276716530971662634, + 0.0137384425896355332126, + 0.00840807615555585383007, + 0.00212825620914618649141, + 0.000250269961544794627958, + 0.113212406648847561139e-4, + }; + static const double q[] = { + 1, + 1.04217814166938418171, + 0.442597659481563127003, + 0.0958492726301061423444, + 0.0105982906484876531489, + 0.000479411269521714493907, + }; + result = y + ff_eval_poly(p, FF_ARRAY_ELEMS(p), z - 3.5) / ff_eval_poly(q, FF_ARRAY_ELEMS(q), z - 3.5); + result *= exp(-z * z) / z; + return 1 - result; + } + /* differ from Boost here, the claim of underflow of erfc(x) past 5.8 is + * slightly incorrect, change to 5.92 + * (really somewhere between 5.9125 and 5.925 is when it saturates) */ + else if (z < 5.92) { + // Max Error found at double precision = 2.997958e-17 + // Maximum Deviation Found: 2.860e-17 + // Expected Error Term: 2.859e-17 + // Maximum Relative Change in Control Points: 1.357e-05 + static const double y = 0.5579090118408203125; + static const double p[] = { + 0.00628057170626964891937, + 0.0175389834052493308818, + -0.212652252872804219852, + -0.687717681153649930619, + -2.5518551727311523996, + -3.22729451764143718517, + -2.8175401114513378771, + }; + static const double q[] = { + 1, + 2.79257750980575282228, + 11.0567237927800161565, + 15.930646027911794143, + 22.9367376522880577224, + 13.5064170191802889145, + 5.48409182238641741584, + }; + result = y + ff_eval_poly(p, FF_ARRAY_ELEMS(p), 1 / z) / ff_eval_poly(q, FF_ARRAY_ELEMS(q), 1 / z); + result *= exp(-z * z) / z; + return 1 - result; + } + /* handle the nan case, but don't use isnan for max portability */ + else if (z != z) + return z; + /* finally return saturated result */ + else + return 1; +} +#endif /* HAVE_ERF */ #if !HAVE_EXPF #undef expf #define expf(x) ((float)exp(x)) -#endif +#endif /* HAVE_EXPF */ #if !HAVE_EXP2 #undef exp2 -#define exp2(x) exp((x) * 0.693147180559945) +#define exp2(x) exp((x) * M_LN2) #endif /* HAVE_EXP2 */ #if !HAVE_EXP2F @@ -72,29 +294,100 @@ static av_always_inline float cbrtf(float x) #endif /* HAVE_EXP2F */ #if !HAVE_ISINF -static av_always_inline av_const int isinf(float x) +#undef isinf +/* Note: these do not follow the BSD/Apple/GNU convention of returning -1 for +-Inf, +1 for Inf, 0 otherwise, but merely follow the POSIX/ISO mandated spec of +returning a non-zero value for +/-Inf, 0 otherwise. */ +static av_always_inline av_const int avpriv_isinff(float x) { uint32_t v = av_float2int(x); if ((v & 0x7f800000) != 0x7f800000) return 0; return !(v & 0x007fffff); } + +static av_always_inline av_const int avpriv_isinf(double x) +{ + uint64_t v = av_double2int(x); + if ((v & 0x7ff0000000000000) != 0x7ff0000000000000) + return 0; + return !(v & 0x000fffffffffffff); +} + +#define isinf(x) \ + (sizeof(x) == sizeof(float) \ + ? avpriv_isinff(x) \ + : avpriv_isinf(x)) #endif /* HAVE_ISINF */ #if !HAVE_ISNAN -static av_always_inline av_const int isnan(float x) +static av_always_inline av_const int avpriv_isnanf(float x) { uint32_t v = av_float2int(x); if ((v & 0x7f800000) != 0x7f800000) return 0; return v & 0x007fffff; } + +static av_always_inline av_const int avpriv_isnan(double x) +{ + uint64_t v = av_double2int(x); + if ((v & 0x7ff0000000000000) != 0x7ff0000000000000) + return 0; + return (v & 0x000fffffffffffff) && 1; +} + +#define isnan(x) \ + (sizeof(x) == sizeof(float) \ + ? avpriv_isnanf(x) \ + : avpriv_isnan(x)) #endif /* HAVE_ISNAN */ +#if !HAVE_ISFINITE +static av_always_inline av_const int avpriv_isfinitef(float x) +{ + uint32_t v = av_float2int(x); + return (v & 0x7f800000) != 0x7f800000; +} + +static av_always_inline av_const int avpriv_isfinite(double x) +{ + uint64_t v = av_double2int(x); + return (v & 0x7ff0000000000000) != 0x7ff0000000000000; +} + +#define isfinite(x) \ + (sizeof(x) == sizeof(float) \ + ? avpriv_isfinitef(x) \ + : avpriv_isfinite(x)) +#endif /* HAVE_ISFINITE */ + +#if !HAVE_HYPOT +static inline av_const double hypot(double x, double y) +{ + double ret, temp; + x = fabs(x); + y = fabs(y); + + if (isinf(x) || isinf(y)) + return av_int2double(0x7ff0000000000000); + if (x == 0 || y == 0) + return x + y; + if (x < y) { + temp = x; + x = y; + y = temp; + } + + y = y/x; + return x*sqrt(1 + y*y); +} +#endif /* HAVE_HYPOT */ + #if !HAVE_LDEXPF #undef ldexpf #define ldexpf(x, exp) ((float)ldexp(x, exp)) -#endif +#endif /* HAVE_LDEXPF */ #if !HAVE_LLRINT #undef llrint @@ -119,12 +412,12 @@ static av_always_inline av_const int isnan(float x) #if !HAVE_LOG10F #undef log10f #define log10f(x) ((float)log10(x)) -#endif +#endif /* HAVE_LOG10F */ #if !HAVE_SINF #undef sinf #define sinf(x) ((float)sin(x)) -#endif +#endif /* HAVE_SINF */ #if !HAVE_RINT static inline double rint(double x) diff --git a/libavutil/lls-test.c b/libavutil/lls-test.c index f9814b3863..12215c5257 100644 --- a/libavutil/lls-test.c +++ b/libavutil/lls-test.c @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/lls.c b/libavutil/lls.c index 97bffb4018..0560b6a79c 100644 --- a/libavutil/lls.c +++ b/libavutil/lls.c @@ -3,20 +3,20 @@ * * Copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -33,7 +33,7 @@ #include "version.h" #include "lls.h" -static void update_lls(LLSModel *m, double *var) +static void update_lls(LLSModel *m, const double *var) { int i, j; @@ -56,7 +56,7 @@ void avpriv_solve_lls(LLSModel *m, double threshold, unsigned short min_order) for (j = i; j < count; j++) { double sum = covar[i][j]; - for (k = i - 1; k >= 0; k--) + for (k = 0; k <= i-1; k++) sum -= factor[i][k] * factor[j][k]; if (i == j) { @@ -72,7 +72,7 @@ void avpriv_solve_lls(LLSModel *m, double threshold, unsigned short min_order) for (i = 0; i < count; i++) { double sum = covar_y[i + 1]; - for (k = i - 1; k >= 0; k--) + for (k = 0; k <= i-1; k++) sum -= factor[i][k] * m->coeff[0][k]; m->coeff[0][i] = sum / factor[i][i]; @@ -101,7 +101,7 @@ void avpriv_solve_lls(LLSModel *m, double threshold, unsigned short min_order) } } -static double evaluate_lls(LLSModel *m, double *param, int order) +static double evaluate_lls(LLSModel *m, const double *param, int order) { int i; double out = 0; diff --git a/libavutil/lls.h b/libavutil/lls.h index 3977e97235..1a276d537d 100644 --- a/libavutil/lls.h +++ b/libavutil/lls.h @@ -3,20 +3,20 @@ * * Copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -47,14 +47,14 @@ typedef struct LLSModel { * 32-byte aligned, and any padding elements must be initialized * (i.e not denormal/nan). */ - void (*update_lls)(struct LLSModel *m, double *var); + void (*update_lls)(struct LLSModel *m, const double *var); /** * Inner product of var[] and the LPC coefs. * @param m this context * @param var training samples, excluding the value to be predicted. unaligned. * @param order lpc order */ - double (*evaluate_lls)(struct LLSModel *m, double *var, int order); + double (*evaluate_lls)(struct LLSModel *m, const double *var, int order); } LLSModel; void avpriv_init_lls(LLSModel *m, int indep_count); diff --git a/libavutil/log-test.c b/libavutil/log-test.c new file mode 100644 index 0000000000..14d476436f --- /dev/null +++ b/libavutil/log-test.c @@ -0,0 +1,71 @@ +/* + * log functions + * Copyright (c) 2003 Michel Bardiaux + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "log.c" + +#include <string.h> + +static int call_log_format_line2(const char *fmt, char *buffer, int buffer_size, ...) +{ + va_list args; + int ret; + int print_prefix=1; + va_start(args, buffer_size); + ret = av_log_format_line2(NULL, AV_LOG_INFO, fmt, args, buffer, buffer_size, &print_prefix); + va_end(args); + return ret; +} + +int main(int argc, char **argv) +{ + int i; + av_log_set_level(AV_LOG_DEBUG); + for (use_color=0; use_color<=256; use_color = 255*use_color+1) { + av_log(NULL, AV_LOG_FATAL, "use_color: %d\n", use_color); + for (i = AV_LOG_DEBUG; i>=AV_LOG_QUIET; i-=8) { + av_log(NULL, i, " %d", i); + av_log(NULL, AV_LOG_INFO, "e "); + av_log(NULL, i + 256*123, "C%d", i); + av_log(NULL, AV_LOG_INFO, "e"); + } + av_log(NULL, AV_LOG_PANIC, "\n"); + } + { + int result; + char buffer[4]; + result = call_log_format_line2("foo", NULL, 0); + if(result != 3) { + printf("Test NULL buffer failed.\n"); + return 1; + } + result = call_log_format_line2("foo", buffer, 2); + if(result != 3 || strncmp(buffer, "f", 2)) { + printf("Test buffer too small failed.\n"); + return 1; + } + result = call_log_format_line2("foo", buffer, 4); + if(result != 3 || strncmp(buffer, "foo", 4)) { + printf("Test buffer sufficiently big failed.\n"); + return 1; + } + } + return 0; +} diff --git a/libavutil/log.c b/libavutil/log.c index 37427efaf1..44c11cb091 100644 --- a/libavutil/log.c +++ b/libavutil/log.c @@ -2,20 +2,20 @@ * log functions * Copyright (c) 2003 Michel Bardiaux * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -34,12 +34,19 @@ #endif #include <stdarg.h> #include <stdlib.h> -#include "avstring.h" #include "avutil.h" +#include "bprint.h" #include "common.h" #include "internal.h" #include "log.h" +#if HAVE_PTHREADS +#include <pthread.h> +static pthread_mutex_t mutex = PTHREAD_MUTEX_INITIALIZER; +#endif + +#define LINE_SZ 1024 + #if HAVE_VALGRIND_VALGRIND_H #include <valgrind/valgrind.h> /* this is the log level at which valgrind will output a full backtrace */ @@ -50,27 +57,74 @@ static int av_log_level = AV_LOG_INFO; static int flags; #define NB_LEVELS 8 -#if HAVE_SETCONSOLETEXTATTRIBUTE +#if defined(_WIN32) && !defined(__MINGW32CE__) && HAVE_SETCONSOLETEXTATTRIBUTE #include <windows.h> -static const uint8_t color[NB_LEVELS] = { 12, 12, 12, 14, 7, 10, 11, 8}; +static const uint8_t color[16 + AV_CLASS_CATEGORY_NB] = { + [AV_LOG_PANIC /8] = 12, + [AV_LOG_FATAL /8] = 12, + [AV_LOG_ERROR /8] = 12, + [AV_LOG_WARNING/8] = 14, + [AV_LOG_INFO /8] = 7, + [AV_LOG_VERBOSE/8] = 10, + [AV_LOG_DEBUG /8] = 10, + [AV_LOG_TRACE /8] = 8, + [16+AV_CLASS_CATEGORY_NA ] = 7, + [16+AV_CLASS_CATEGORY_INPUT ] = 13, + [16+AV_CLASS_CATEGORY_OUTPUT ] = 5, + [16+AV_CLASS_CATEGORY_MUXER ] = 13, + [16+AV_CLASS_CATEGORY_DEMUXER ] = 5, + [16+AV_CLASS_CATEGORY_ENCODER ] = 11, + [16+AV_CLASS_CATEGORY_DECODER ] = 3, + [16+AV_CLASS_CATEGORY_FILTER ] = 10, + [16+AV_CLASS_CATEGORY_BITSTREAM_FILTER] = 9, + [16+AV_CLASS_CATEGORY_SWSCALER ] = 7, + [16+AV_CLASS_CATEGORY_SWRESAMPLER ] = 7, + [16+AV_CLASS_CATEGORY_DEVICE_VIDEO_OUTPUT ] = 13, + [16+AV_CLASS_CATEGORY_DEVICE_VIDEO_INPUT ] = 5, + [16+AV_CLASS_CATEGORY_DEVICE_AUDIO_OUTPUT ] = 13, + [16+AV_CLASS_CATEGORY_DEVICE_AUDIO_INPUT ] = 5, + [16+AV_CLASS_CATEGORY_DEVICE_OUTPUT ] = 13, + [16+AV_CLASS_CATEGORY_DEVICE_INPUT ] = 5, +}; + static int16_t background, attr_orig; static HANDLE con; -#define set_color(x) SetConsoleTextAttribute(con, background | color[x]) -#define reset_color() SetConsoleTextAttribute(con, attr_orig) -#define print_256color(x) #else -static const uint8_t color[NB_LEVELS] = { - 0x41, 0x41, 0x11, 0x03, 9, 0x02, 0x06, 0x07 + +static const uint32_t color[16 + AV_CLASS_CATEGORY_NB] = { + [AV_LOG_PANIC /8] = 52 << 16 | 196 << 8 | 0x41, + [AV_LOG_FATAL /8] = 208 << 8 | 0x41, + [AV_LOG_ERROR /8] = 196 << 8 | 0x11, + [AV_LOG_WARNING/8] = 226 << 8 | 0x03, + [AV_LOG_INFO /8] = 253 << 8 | 0x09, + [AV_LOG_VERBOSE/8] = 40 << 8 | 0x02, + [AV_LOG_DEBUG /8] = 34 << 8 | 0x02, + [AV_LOG_TRACE /8] = 34 << 8 | 0x07, + [16+AV_CLASS_CATEGORY_NA ] = 250 << 8 | 0x09, + [16+AV_CLASS_CATEGORY_INPUT ] = 219 << 8 | 0x15, + [16+AV_CLASS_CATEGORY_OUTPUT ] = 201 << 8 | 0x05, + [16+AV_CLASS_CATEGORY_MUXER ] = 213 << 8 | 0x15, + [16+AV_CLASS_CATEGORY_DEMUXER ] = 207 << 8 | 0x05, + [16+AV_CLASS_CATEGORY_ENCODER ] = 51 << 8 | 0x16, + [16+AV_CLASS_CATEGORY_DECODER ] = 39 << 8 | 0x06, + [16+AV_CLASS_CATEGORY_FILTER ] = 155 << 8 | 0x12, + [16+AV_CLASS_CATEGORY_BITSTREAM_FILTER] = 192 << 8 | 0x14, + [16+AV_CLASS_CATEGORY_SWSCALER ] = 153 << 8 | 0x14, + [16+AV_CLASS_CATEGORY_SWRESAMPLER ] = 147 << 8 | 0x14, + [16+AV_CLASS_CATEGORY_DEVICE_VIDEO_OUTPUT ] = 213 << 8 | 0x15, + [16+AV_CLASS_CATEGORY_DEVICE_VIDEO_INPUT ] = 207 << 8 | 0x05, + [16+AV_CLASS_CATEGORY_DEVICE_AUDIO_OUTPUT ] = 213 << 8 | 0x15, + [16+AV_CLASS_CATEGORY_DEVICE_AUDIO_INPUT ] = 207 << 8 | 0x05, + [16+AV_CLASS_CATEGORY_DEVICE_OUTPUT ] = 213 << 8 | 0x15, + [16+AV_CLASS_CATEGORY_DEVICE_INPUT ] = 207 << 8 | 0x05, }; -#define set_color(x) fprintf(stderr, "\033[%d;3%dm", color[x] >> 4, color[x]&15) -#define print_256color(x) fprintf(stderr, "\033[38;5;%dm", x) -#define reset_color() fprintf(stderr, "\033[0m") + #endif static int use_color = -1; static void check_color_terminal(void) { -#if HAVE_SETCONSOLETEXTATTRIBUTE +#if defined(_WIN32) && !defined(__MINGW32CE__) && HAVE_SETCONSOLETEXTATTRIBUTE CONSOLE_SCREEN_BUFFER_INFO con_info; con = GetStdHandle(STD_ERROR_HANDLE); use_color = (con != INVALID_HANDLE_VALUE) && !getenv("NO_COLOR") && @@ -84,8 +138,9 @@ static void check_color_terminal(void) char *term = getenv("TERM"); use_color = !getenv("NO_COLOR") && !getenv("AV_LOG_FORCE_NOCOLOR") && (getenv("TERM") && isatty(2) || getenv("AV_LOG_FORCE_COLOR")); - if (use_color) - use_color += term && strstr(term, "256color"); + if ( getenv("AV_LOG_FORCE_256COLOR") + || (term && strstr(term, "256color"))) + use_color *= 256; #else use_color = getenv("AV_LOG_FORCE_COLOR") && !getenv("NO_COLOR") && !getenv("AV_LOG_FORCE_NOCOLOR"); @@ -94,25 +149,45 @@ static void check_color_terminal(void) static void colored_fputs(int level, int tint, const char *str) { + int local_use_color; + if (!*str) + return; + if (use_color < 0) check_color_terminal(); - switch (use_color) { - case 1: - set_color(level); - break; - case 2: - set_color(level); - if (tint) - print_256color(tint); - break; - default: - break; - } + if (level == AV_LOG_INFO/8) local_use_color = 0; + else local_use_color = use_color; + +#if defined(_WIN32) && !defined(__MINGW32CE__) && HAVE_SETCONSOLETEXTATTRIBUTE + if (local_use_color) + SetConsoleTextAttribute(con, background | color[level]); fputs(str, stderr); - if (use_color) { - reset_color(); - } + if (local_use_color) + SetConsoleTextAttribute(con, attr_orig); +#else + if (local_use_color == 1) { + fprintf(stderr, + "\033[%d;3%dm%s\033[0m", + (color[level] >> 4) & 15, + color[level] & 15, + str); + } else if (tint && use_color == 256) { + fprintf(stderr, + "\033[48;5;%dm\033[38;5;%dm%s\033[0m", + (color[level] >> 16) & 0xff, + tint, + str); + } else if (local_use_color == 256) { + fprintf(stderr, + "\033[48;5;%dm\033[38;5;%dm%s\033[0m", + (color[level] >> 16) & 0xff, + (color[level] >> 8) & 0xff, + str); + } else + fputs(str, stderr); +#endif + } const char *av_default_item_name(void *ptr) @@ -120,60 +195,169 @@ const char *av_default_item_name(void *ptr) return (*(AVClass **) ptr)->class_name; } -void av_log_default_callback(void *avcl, int level, const char *fmt, va_list vl) +AVClassCategory av_default_get_category(void *ptr) { - static int print_prefix = 1; - static int count; - static char prev[1024]; - char line[1024]; - static int is_atty; - AVClass* avc = avcl ? *(AVClass **) avcl : NULL; - unsigned tint = level & 0xff00; + return (*(AVClass **) ptr)->category; +} - level &= 0xff; +static void sanitize(uint8_t *line){ + while(*line){ + if(*line < 0x08 || (*line > 0x0D && *line < 0x20)) + *line='?'; + line++; + } +} - if (level > av_log_level) - return; - line[0] = 0; - if (print_prefix && avc) { +static int get_category(void *ptr){ + AVClass *avc = *(AVClass **) ptr; + if( !avc + || (avc->version&0xFF)<100 + || avc->version < (51 << 16 | 59 << 8) + || avc->category >= AV_CLASS_CATEGORY_NB) return AV_CLASS_CATEGORY_NA + 16; + + if(avc->get_category) + return avc->get_category(ptr) + 16; + + return avc->category + 16; +} + +static const char *get_level_str(int level) +{ + switch (level) { + case AV_LOG_QUIET: + return "quiet"; + case AV_LOG_DEBUG: + return "debug"; + case AV_LOG_VERBOSE: + return "verbose"; + case AV_LOG_INFO: + return "info"; + case AV_LOG_WARNING: + return "warning"; + case AV_LOG_ERROR: + return "error"; + case AV_LOG_FATAL: + return "fatal"; + case AV_LOG_PANIC: + return "panic"; + default: + return ""; + } +} + +static void format_line(void *avcl, int level, const char *fmt, va_list vl, + AVBPrint part[4], int *print_prefix, int type[2]) +{ + AVClass* avc = avcl ? *(AVClass **) avcl : NULL; + av_bprint_init(part+0, 0, 1); + av_bprint_init(part+1, 0, 1); + av_bprint_init(part+2, 0, 1); + av_bprint_init(part+3, 0, 65536); + + if(type) type[0] = type[1] = AV_CLASS_CATEGORY_NA + 16; + if (*print_prefix && avc) { if (avc->parent_log_context_offset) { AVClass** parent = *(AVClass ***) (((uint8_t *) avcl) + avc->parent_log_context_offset); if (parent && *parent) { - snprintf(line, sizeof(line), "[%s @ %p] ", + av_bprintf(part+0, "[%s @ %p] ", (*parent)->item_name(parent), parent); + if(type) type[0] = get_category(parent); } } - snprintf(line + strlen(line), sizeof(line) - strlen(line), "[%s @ %p] ", + av_bprintf(part+1, "[%s @ %p] ", avc->item_name(avcl), avcl); + if(type) type[1] = get_category(avcl); + + if (flags & AV_LOG_PRINT_LEVEL) + av_bprintf(part+2, "[%s] ", get_level_str(level)); + } + + av_vbprintf(part+3, fmt, vl); + + if(*part[0].str || *part[1].str || *part[2].str || *part[3].str) { + char lastc = part[3].len && part[3].len <= part[3].size ? part[3].str[part[3].len - 1] : 0; + *print_prefix = lastc == '\n' || lastc == '\r'; + } +} + +void av_log_format_line(void *ptr, int level, const char *fmt, va_list vl, + char *line, int line_size, int *print_prefix) +{ + av_log_format_line2(ptr, level, fmt, vl, line, line_size, print_prefix); +} + +int av_log_format_line2(void *ptr, int level, const char *fmt, va_list vl, + char *line, int line_size, int *print_prefix) +{ + AVBPrint part[4]; + int ret; + + format_line(ptr, level, fmt, vl, part, print_prefix, NULL); + ret = snprintf(line, line_size, "%s%s%s%s", part[0].str, part[1].str, part[2].str, part[3].str); + av_bprint_finalize(part+3, NULL); + return ret; +} + +void av_log_default_callback(void* ptr, int level, const char* fmt, va_list vl) +{ + static int print_prefix = 1; + static int count; + static char prev[LINE_SZ]; + AVBPrint part[4]; + char line[LINE_SZ]; + static int is_atty; + int type[2]; + unsigned tint = 0; + + if (level >= 0) { + tint = level & 0xff00; + level &= 0xff; } - vsnprintf(line + strlen(line), sizeof(line) - strlen(line), fmt, vl); + if (level > av_log_level) + return; +#if HAVE_PTHREADS + pthread_mutex_lock(&mutex); +#endif - print_prefix = strlen(line) && line[strlen(line) - 1] == '\n'; + format_line(ptr, level, fmt, vl, part, &print_prefix, type); + snprintf(line, sizeof(line), "%s%s%s%s", part[0].str, part[1].str, part[2].str, part[3].str); #if HAVE_ISATTY if (!is_atty) is_atty = isatty(2) ? 1 : -1; #endif - if (print_prefix && (flags & AV_LOG_SKIP_REPEATED) && - !strncmp(line, prev, sizeof line)) { + if (print_prefix && (flags & AV_LOG_SKIP_REPEATED) && !strcmp(line, prev) && + *line && line[strlen(line) - 1] != '\r'){ count++; if (is_atty == 1) fprintf(stderr, " Last message repeated %d times\r", count); - return; + goto end; } if (count > 0) { fprintf(stderr, " Last message repeated %d times\n", count); count = 0; } - colored_fputs(av_clip(level >> 3, 0, NB_LEVELS - 1), tint >> 8, line); - av_strlcpy(prev, line, sizeof line); + strcpy(prev, line); + sanitize(part[0].str); + colored_fputs(type[0], 0, part[0].str); + sanitize(part[1].str); + colored_fputs(type[1], 0, part[1].str); + sanitize(part[2].str); + colored_fputs(av_clip(level >> 3, 0, NB_LEVELS - 1), tint >> 8, part[2].str); + sanitize(part[3].str); + colored_fputs(av_clip(level >> 3, 0, NB_LEVELS - 1), tint >> 8, part[3].str); #if CONFIG_VALGRIND_BACKTRACE if (level <= BACKTRACE_LOGLEVEL) - VALGRIND_PRINTF_BACKTRACE(""); + VALGRIND_PRINTF_BACKTRACE("%s", ""); +#endif +end: + av_bprint_finalize(part+3, NULL); +#if HAVE_PTHREADS + pthread_mutex_unlock(&mutex); #endif } @@ -194,7 +378,9 @@ void av_log(void* avcl, int level, const char *fmt, ...) void av_vlog(void* avcl, int level, const char *fmt, va_list vl) { - av_log_callback(avcl, level, fmt, vl); + void (*log_callback)(void*, int, const char*, va_list) = av_log_callback; + if (log_callback) + log_callback(avcl, level, fmt, vl); } int av_log_get_level(void) @@ -212,6 +398,11 @@ void av_log_set_flags(int arg) flags = arg; } +int av_log_get_flags(void) +{ + return flags; +} + void av_log_set_callback(void (*callback)(void*, int, const char*, va_list)) { av_log_callback = callback; @@ -221,14 +412,14 @@ static void missing_feature_sample(int sample, void *avc, const char *msg, va_list argument_list) { av_vlog(avc, AV_LOG_WARNING, msg, argument_list); - av_log(avc, AV_LOG_WARNING, " is not implemented. Update your Libav " + av_log(avc, AV_LOG_WARNING, " is not implemented. Update your FFmpeg " "version to the newest one from Git. If the problem still " "occurs, it means that your file has a feature which has not " "been implemented.\n"); if (sample) av_log(avc, AV_LOG_WARNING, "If you want to help, upload a sample " - "of this file to ftp://upload.libav.org/incoming/ " - "and contact the libav-devel mailing list.\n"); + "of this file to ftp://upload.ffmpeg.org/incoming/ " + "and contact the ffmpeg-devel mailing list. (ffmpeg-devel@ffmpeg.org)\n"); } void avpriv_request_sample(void *avc, const char *msg, ...) diff --git a/libavutil/log.h b/libavutil/log.h index d4daea9c1a..0acc1b9214 100644 --- a/libavutil/log.h +++ b/libavutil/log.h @@ -1,20 +1,20 @@ /* * copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -26,6 +26,39 @@ #include "attributes.h" #include "version.h" +typedef enum { + AV_CLASS_CATEGORY_NA = 0, + AV_CLASS_CATEGORY_INPUT, + AV_CLASS_CATEGORY_OUTPUT, + AV_CLASS_CATEGORY_MUXER, + AV_CLASS_CATEGORY_DEMUXER, + AV_CLASS_CATEGORY_ENCODER, + AV_CLASS_CATEGORY_DECODER, + AV_CLASS_CATEGORY_FILTER, + AV_CLASS_CATEGORY_BITSTREAM_FILTER, + AV_CLASS_CATEGORY_SWSCALER, + AV_CLASS_CATEGORY_SWRESAMPLER, + AV_CLASS_CATEGORY_DEVICE_VIDEO_OUTPUT = 40, + AV_CLASS_CATEGORY_DEVICE_VIDEO_INPUT, + AV_CLASS_CATEGORY_DEVICE_AUDIO_OUTPUT, + AV_CLASS_CATEGORY_DEVICE_AUDIO_INPUT, + AV_CLASS_CATEGORY_DEVICE_OUTPUT, + AV_CLASS_CATEGORY_DEVICE_INPUT, + AV_CLASS_CATEGORY_NB, ///< not part of ABI/API +}AVClassCategory; + +#define AV_IS_INPUT_DEVICE(category) \ + (((category) == AV_CLASS_CATEGORY_DEVICE_VIDEO_INPUT) || \ + ((category) == AV_CLASS_CATEGORY_DEVICE_AUDIO_INPUT) || \ + ((category) == AV_CLASS_CATEGORY_DEVICE_INPUT)) + +#define AV_IS_OUTPUT_DEVICE(category) \ + (((category) == AV_CLASS_CATEGORY_DEVICE_VIDEO_OUTPUT) || \ + ((category) == AV_CLASS_CATEGORY_DEVICE_AUDIO_OUTPUT) || \ + ((category) == AV_CLASS_CATEGORY_DEVICE_OUTPUT)) + +struct AVOptionRanges; + /** * Describe the class of an AVClass context structure. That is an * arbitrary struct of which the first field is a pointer to an @@ -88,6 +121,25 @@ typedef struct AVClass { * child_class_next iterates over _all possible_ children. */ const struct AVClass* (*child_class_next)(const struct AVClass *prev); + + /** + * Category used for visualization (like color) + * This is only set if the category is equal for all objects using this class. + * available since version (51 << 16 | 56 << 8 | 100) + */ + AVClassCategory category; + + /** + * Callback to return the category. + * available since version (51 << 16 | 59 << 8 | 100) + */ + AVClassCategory (*get_category)(void* ctx); + + /** + * Callback to return the supported/allowed ranges. + * available since version (52.12) + */ + int (*query_ranges)(struct AVOptionRanges **, void *obj, const char *key, int flags); } AVClass; /** @@ -149,6 +201,8 @@ typedef struct AVClass { */ #define AV_LOG_TRACE 56 +#define AV_LOG_MAX_OFFSET (AV_LOG_TRACE - AV_LOG_QUIET) + /** * @} */ @@ -161,7 +215,7 @@ typedef struct AVClass { * Requires 256color terminal support. Uses outside debugging is not * recommended. */ -#define AV_LOG_C(x) (x << 8) +#define AV_LOG_C(x) ((x) << 8) /** * Send the specified message to the log if the level is less than or equal @@ -171,7 +225,7 @@ typedef struct AVClass { * @see av_log_set_callback * * @param avcl A pointer to an arbitrary struct of which the first field is a - * pointer to an AVClass struct. + * pointer to an AVClass struct or NULL if general log. * @param level The importance level of the message expressed using a @ref * lavu_log_constants "Logging Constant". * @param fmt The format string (printf-compatible) that specifies how @@ -218,6 +272,9 @@ void av_log_set_level(int level); /** * Set the logging callback * + * @note The callback must be thread safe, even if the application does not use + * threads itself as some codecs are multithreaded. + * * @see av_log_default_callback * * @param callback A logging function with a compatible signature. @@ -248,6 +305,34 @@ void av_log_default_callback(void *avcl, int level, const char *fmt, * @return The AVClass class_name */ const char* av_default_item_name(void* ctx); +AVClassCategory av_default_get_category(void *ptr); + +/** + * Format a line of log the same way as the default callback. + * @param line buffer to receive the formatted line + * @param line_size size of the buffer + * @param print_prefix used to store whether the prefix must be printed; + * must point to a persistent integer initially set to 1 + */ +void av_log_format_line(void *ptr, int level, const char *fmt, va_list vl, + char *line, int line_size, int *print_prefix); + +/** + * Format a line of log the same way as the default callback. + * @param line buffer to receive the formatted line; + * may be NULL if line_size is 0 + * @param line_size size of the buffer; at most line_size-1 characters will + * be written to the buffer, plus one null terminator + * @param print_prefix used to store whether the prefix must be printed; + * must point to a persistent integer initially set to 1 + * @return Returns a negative value if an error occurred, otherwise returns + * the number of characters that would have been written for a + * sufficiently large buffer, not including the terminating null + * character. If the return value is not less than line_size, it means + * that the log message was truncated to fit the buffer. + */ +int av_log_format_line2(void *ptr, int level, const char *fmt, va_list vl, + char *line, int line_size, int *print_prefix); #if FF_API_DLOG /** @@ -259,7 +344,7 @@ const char* av_default_item_name(void* ctx); #ifdef DEBUG # define av_dlog(pctx, ...) av_log(pctx, AV_LOG_DEBUG, __VA_ARGS__) #else -# define av_dlog(pctx, ...) +# define av_dlog(pctx, ...) do { if (0) av_log(pctx, AV_LOG_DEBUG, __VA_ARGS__); } while (0) #endif #endif /* FF_API_DLOG */ @@ -269,10 +354,20 @@ const char* av_default_item_name(void* ctx); * "Last message repeated x times" messages below (f)printf messages with some * bad luck. * Also to receive the last, "last repeated" line if any, the user app must - * call av_log(NULL, AV_LOG_QUIET, ""); at the end + * call av_log(NULL, AV_LOG_QUIET, "%s", ""); at the end */ #define AV_LOG_SKIP_REPEATED 1 + +/** + * Include the log severity in messages originating from codecs. + * + * Results in messages such as: + * [rawvideo @ 0xDEADBEEF] [error] encode did not produce valid pts + */ +#define AV_LOG_PRINT_LEVEL 2 + void av_log_set_flags(int arg); +int av_log_get_flags(void); /** * @} diff --git a/libavutil/log2_tab.c b/libavutil/log2_tab.c index f6cbe79063..0dbf07d74c 100644 --- a/libavutil/log2_tab.c +++ b/libavutil/log2_tab.c @@ -1,18 +1,20 @@ /* - * This file is part of Libav. + * Copyright (c) 2003-2012 Michael Niedermayer <michaelni@gmx.at> * - * Libav is free software; you can redistribute it and/or + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/lzo.c b/libavutil/lzo.c index e458165261..bca10ec3e0 100644 --- a/libavutil/lzo.c +++ b/libavutil/lzo.c @@ -2,26 +2,27 @@ * LZO 1x decompression * Copyright (c) 2006 Reimar Doeffinger * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ #include <string.h> #include "avutil.h" +#include "avassert.h" #include "common.h" #include "intreadwrite.h" #include "lzo.h" @@ -65,8 +66,13 @@ static inline int get_len(LZOContext *c, int x, int mask) { int cnt = x & mask; if (!cnt) { - while (!(x = get_byte(c))) + while (!(x = get_byte(c))) { + if (cnt >= INT_MAX - 1000) { + c->error |= AV_LZO_ERROR; + break; + } cnt += 255; + } cnt += mask + x; } return cnt; @@ -80,10 +86,7 @@ static inline void copy(LZOContext *c, int cnt) { register const uint8_t *src = c->in; register uint8_t *dst = c->out; - if (cnt < 0) { - c->error |= AV_LZO_ERROR; - return; - } + av_assert0(cnt >= 0); if (cnt > c->in_end - src) { cnt = FFMAX(c->in_end - src, 0); c->error |= AV_LZO_INPUT_DEPLETED; @@ -106,7 +109,7 @@ static inline void copy(LZOContext *c, int cnt) /** * @brief Copies previously decoded bytes to current position. - * @param back how many bytes back we start + * @param back how many bytes back we start, must be > 0 * @param cnt number of bytes to copy, must be > 0 * * cnt > back is valid, this will copy the bytes we just copied, @@ -115,10 +118,7 @@ static inline void copy(LZOContext *c, int cnt) static inline void copy_backptr(LZOContext *c, int back, int cnt) { register uint8_t *dst = c->out; - if (cnt <= 0) { - c->error |= AV_LZO_ERROR; - return; - } + av_assert0(cnt > 0); if (dst - c->out_start < back) { c->error |= AV_LZO_INVALID_BACKPTR; return; @@ -136,11 +136,11 @@ int av_lzo1x_decode(void *out, int *outlen, const void *in, int *inlen) int state = 0; int x; LZOContext c; - if (!*outlen || !*inlen) { + if (*outlen <= 0 || *inlen <= 0) { int res = 0; - if (!*outlen) + if (*outlen <= 0) res |= AV_LZO_OUTPUT_FULL; - if (!*inlen) + if (*inlen <= 0) res |= AV_LZO_INPUT_DEPLETED; return res; } @@ -203,3 +203,58 @@ int av_lzo1x_decode(void *out, int *outlen, const void *in, int *inlen) *outlen = c.out_end - c.out; return c.error; } + +#ifdef TEST +#include <stdio.h> +#include <lzo/lzo1x.h> +#include "log.h" +#define MAXSZ (10*1024*1024) + +/* Define one of these to 1 if you wish to benchmark liblzo + * instead of our native implementation. */ +#define BENCHMARK_LIBLZO_SAFE 0 +#define BENCHMARK_LIBLZO_UNSAFE 0 + +int main(int argc, char *argv[]) { + FILE *in = fopen(argv[1], "rb"); + int comp_level = argc > 2 ? atoi(argv[2]) : 0; + uint8_t *orig = av_malloc(MAXSZ + 16); + uint8_t *comp = av_malloc(2*MAXSZ + 16); + uint8_t *decomp = av_malloc(MAXSZ + 16); + size_t s = fread(orig, 1, MAXSZ, in); + lzo_uint clen = 0; + long tmp[LZO1X_MEM_COMPRESS]; + int inlen, outlen; + int i; + av_log_set_level(AV_LOG_DEBUG); + if (comp_level == 0) { + lzo1x_1_compress(orig, s, comp, &clen, tmp); + } else if (comp_level == 11) { + lzo1x_1_11_compress(orig, s, comp, &clen, tmp); + } else if (comp_level == 12) { + lzo1x_1_12_compress(orig, s, comp, &clen, tmp); + } else if (comp_level == 15) { + lzo1x_1_15_compress(orig, s, comp, &clen, tmp); + } else + lzo1x_999_compress(orig, s, comp, &clen, tmp); + for (i = 0; i < 300; i++) { +START_TIMER + inlen = clen; outlen = MAXSZ; +#if BENCHMARK_LIBLZO_SAFE + if (lzo1x_decompress_safe(comp, inlen, decomp, &outlen, NULL)) +#elif BENCHMARK_LIBLZO_UNSAFE + if (lzo1x_decompress(comp, inlen, decomp, &outlen, NULL)) +#else + if (av_lzo1x_decode(decomp, &outlen, comp, &inlen)) +#endif + av_log(NULL, AV_LOG_ERROR, "decompression error\n"); +STOP_TIMER("lzod") + } + if (memcmp(orig, decomp, s)) + av_log(NULL, AV_LOG_ERROR, "decompression incorrect\n"); + else + av_log(NULL, AV_LOG_ERROR, "decompression OK\n"); + fclose(in); + return 0; +} +#endif diff --git a/libavutil/lzo.h b/libavutil/lzo.h index 9d7e8f1dc1..c03403992d 100644 --- a/libavutil/lzo.h +++ b/libavutil/lzo.h @@ -2,20 +2,20 @@ * LZO 1x decompression * copyright (c) 2006 Reimar Doeffinger * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/macros.h b/libavutil/macros.h index 3e7b00559f..2007ee5619 100644 --- a/libavutil/macros.h +++ b/libavutil/macros.h @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/mastering_display_metadata.c b/libavutil/mastering_display_metadata.c new file mode 100644 index 0000000000..e1683e55c7 --- /dev/null +++ b/libavutil/mastering_display_metadata.c @@ -0,0 +1,43 @@ +/** + * Copyright (c) 2016 Neil Birkbeck <neil.birkbeck@gmail.com> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include <stdint.h> +#include <string.h> + +#include "mastering_display_metadata.h" +#include "mem.h" + +AVMasteringDisplayMetadata *av_mastering_display_metadata_alloc(void) +{ + return av_mallocz(sizeof(AVMasteringDisplayMetadata)); +} + +AVMasteringDisplayMetadata *av_mastering_display_metadata_create_side_data(AVFrame *frame) +{ + AVFrameSideData *side_data = av_frame_new_side_data(frame, + AV_FRAME_DATA_MASTERING_DISPLAY_METADATA, + sizeof(AVMasteringDisplayMetadata)); + if (!side_data) + return NULL; + + memset(side_data->data, 0, sizeof(AVMasteringDisplayMetadata)); + + return (AVMasteringDisplayMetadata *)side_data->data; +} diff --git a/libavutil/mastering_display_metadata.h b/libavutil/mastering_display_metadata.h new file mode 100644 index 0000000000..936533fec4 --- /dev/null +++ b/libavutil/mastering_display_metadata.h @@ -0,0 +1,89 @@ +/** + * Copyright (c) 2016 Neil Birkbeck <neil.birkbeck@gmail.com> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#ifndef AVUTIL_MASTERING_DISPLAY_METADATA_H +#define AVUTIL_MASTERING_DISPLAY_METADATA_H + +#include "frame.h" +#include "rational.h" + + +/** + * Mastering display metadata capable of representing the color volume of + * the display used to master the content (SMPTE 2086:2014). + * + * To be used as payload of a AVFrameSideData or AVPacketSideData with the + * appropriate type. + * + * @note The struct should be allocated with av_mastering_display_metadata_alloc() + * and its size is not a part of the public ABI. + */ +typedef struct AVMasteringDisplayMetadata { + /** + * CIE 1931 xy chromaticity coords of color primaries (r, g, b order). + */ + AVRational display_primaries[3][2]; + + /** + * CIE 1931 xy chromaticity coords of white point. + */ + AVRational white_point[2]; + + /** + * Min luminance of mastering display (cd/m^2). + */ + AVRational min_luminance; + + /** + * Max luminance of mastering display (cd/m^2). + */ + AVRational max_luminance; + + /** + * Flag indicating whether the display primaries (and white point) are set. + */ + int has_primaries; + + /** + * Flag indicating whether the luminance (min_ and max_) have been set. + */ + int has_luminance; + +} AVMasteringDisplayMetadata; + +/** + * Allocate an AVMasteringDisplayMetadata structure and set its fields to + * default values. The resulting struct can be freed using av_freep(). + * + * @return An AVMasteringDisplayMetadata filled with default values or NULL + * on failure. + */ +AVMasteringDisplayMetadata *av_mastering_display_metadata_alloc(void); + +/** + * Allocate a complete AVMasteringDisplayMetadata and add it to the frame. + * + * @param frame The frame which side data is added to. + * + * @return The AVMasteringDisplayMetadata structure to be filled by caller. + */ +AVMasteringDisplayMetadata *av_mastering_display_metadata_create_side_data(AVFrame *frame); + +#endif /* AVUTIL_MASTERING_DISPLAY_METADATA_H */ diff --git a/libavutil/mathematics.c b/libavutil/mathematics.c index c9b24afeac..20ff37f5e9 100644 --- a/libavutil/mathematics.c +++ b/libavutil/mathematics.c @@ -1,20 +1,20 @@ /* - * Copyright (c) 2005 Michael Niedermayer <michaelni@gmx.at> + * Copyright (c) 2005-2012 Michael Niedermayer <michaelni@gmx.at> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -27,25 +27,52 @@ #include <limits.h> #include "mathematics.h" +#include "libavutil/intmath.h" +#include "libavutil/common.h" +#include "avassert.h" #include "version.h" -int64_t av_gcd(int64_t a, int64_t b) -{ - if (b) - return av_gcd(b, a % b); - else +/* Stein's binary GCD algorithm: + * https://en.wikipedia.org/wiki/Binary_GCD_algorithm */ +int64_t av_gcd(int64_t a, int64_t b) { + int za, zb, k; + int64_t u, v; + if (a == 0) + return b; + if (b == 0) return a; + za = ff_ctzll(a); + zb = ff_ctzll(b); + k = FFMIN(za, zb); + u = llabs(a >> za); + v = llabs(b >> zb); + while (u != v) { + if (u > v) + FFSWAP(int64_t, v, u); + v -= u; + v >>= ff_ctzll(v); + } + return (uint64_t)u << k; } int64_t av_rescale_rnd(int64_t a, int64_t b, int64_t c, enum AVRounding rnd) { int64_t r = 0; + av_assert2(c > 0); + av_assert2(b >=0); + av_assert2((unsigned)(rnd&~AV_ROUND_PASS_MINMAX)<=5 && (rnd&~AV_ROUND_PASS_MINMAX)!=4); - if (c <= 0 || b < 0 || rnd == 4 || rnd > 5) + if (c <= 0 || b < 0 || !((unsigned)(rnd&~AV_ROUND_PASS_MINMAX)<=5 && (rnd&~AV_ROUND_PASS_MINMAX)!=4)) return INT64_MIN; - if (a < 0 && a != INT64_MIN) - return -av_rescale_rnd(-a, b, c, rnd ^ ((rnd >> 1) & 1)); + if (rnd & AV_ROUND_PASS_MINMAX) { + if (a == INT64_MIN || a == INT64_MAX) + return a; + rnd -= AV_ROUND_PASS_MINMAX; + } + + if (a < 0) + return -(uint64_t)av_rescale_rnd(-FFMAX(a, -INT64_MAX), b, c, rnd ^ ((rnd >> 1) & 1)); if (rnd == AV_ROUND_NEAR_INF) r = c / 2; @@ -55,8 +82,13 @@ int64_t av_rescale_rnd(int64_t a, int64_t b, int64_t c, enum AVRounding rnd) if (b <= INT_MAX && c <= INT_MAX) { if (a <= INT_MAX) return (a * b + r) / c; - else - return a / c * b + (a % c * b + r) / c; + else { + int64_t ad = a / c; + int64_t a2 = (a % c * b + r) / c; + if (ad >= INT32_MAX && b && ad > (INT64_MAX - a2) / b) + return INT64_MIN; + return ad * b + a2; + } } else { #if 1 uint64_t a0 = a & 0xFFFFFFFF; @@ -80,6 +112,8 @@ int64_t av_rescale_rnd(int64_t a, int64_t b, int64_t c, enum AVRounding rnd) t1++; } } + if (t1 > INT64_MAX) + return INT64_MIN; return t1; } #else @@ -114,6 +148,8 @@ int av_compare_ts(int64_t ts_a, AVRational tb_a, int64_t ts_b, AVRational tb_b) { int64_t a = tb_a.num * (int64_t)tb_b.den; int64_t b = tb_b.num * (int64_t)tb_a.den; + if ((FFABS(ts_a)|a|FFABS(ts_b)|b) <= INT_MAX) + return (ts_a*a > ts_b*b) - (ts_a*a < ts_b*b); if (av_rescale_rnd(ts_a, a, b, AV_ROUND_DOWN) < ts_b) return -1; if (av_rescale_rnd(ts_b, b, a, AV_ROUND_DOWN) < ts_a) @@ -128,3 +164,48 @@ int64_t av_compare_mod(uint64_t a, uint64_t b, uint64_t mod) c -= mod; return c; } + +int64_t av_rescale_delta(AVRational in_tb, int64_t in_ts, AVRational fs_tb, int duration, int64_t *last, AVRational out_tb){ + int64_t a, b, this; + + av_assert0(in_ts != AV_NOPTS_VALUE); + av_assert0(duration >= 0); + + if (*last == AV_NOPTS_VALUE || !duration || in_tb.num*(int64_t)out_tb.den <= out_tb.num*(int64_t)in_tb.den) { +simple_round: + *last = av_rescale_q(in_ts, in_tb, fs_tb) + duration; + return av_rescale_q(in_ts, in_tb, out_tb); + } + + a = av_rescale_q_rnd(2*in_ts-1, in_tb, fs_tb, AV_ROUND_DOWN) >>1; + b = (av_rescale_q_rnd(2*in_ts+1, in_tb, fs_tb, AV_ROUND_UP )+1)>>1; + if (*last < 2*a - b || *last > 2*b - a) + goto simple_round; + + this = av_clip64(*last, a, b); + *last = this + duration; + + return av_rescale_q(this, fs_tb, out_tb); +} + +int64_t av_add_stable(AVRational ts_tb, int64_t ts, AVRational inc_tb, int64_t inc) +{ + int64_t m, d; + + if (inc != 1) + inc_tb = av_mul_q(inc_tb, (AVRational) {inc, 1}); + + m = inc_tb.num * (int64_t)ts_tb.den; + d = inc_tb.den * (int64_t)ts_tb.num; + + if (m % d == 0) + return ts + m / d; + if (m < d) + return ts; + + { + int64_t old = av_rescale_q(ts, ts_tb, inc_tb); + int64_t old_ts = av_rescale_q(old, inc_tb, ts_tb); + return av_rescale_q(old + 1, inc_tb, ts_tb) + (ts - old_ts); + } +} diff --git a/libavutil/mathematics.h b/libavutil/mathematics.h index 043dd0fafe..57c44f845d 100644 --- a/libavutil/mathematics.h +++ b/libavutil/mathematics.h @@ -1,20 +1,20 @@ /* - * copyright (c) 2005 Michael Niedermayer <michaelni@gmx.at> + * copyright (c) 2005-2012 Michael Niedermayer <michaelni@gmx.at> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -27,12 +27,33 @@ #include "rational.h" #include "intfloat.h" +#ifndef M_E +#define M_E 2.7182818284590452354 /* e */ +#endif +#ifndef M_LN2 +#define M_LN2 0.69314718055994530942 /* log_e 2 */ +#endif +#ifndef M_LN10 +#define M_LN10 2.30258509299404568402 /* log_e 10 */ +#endif #ifndef M_LOG2_10 #define M_LOG2_10 3.32192809488736234787 /* log_2 10 */ #endif #ifndef M_PHI #define M_PHI 1.61803398874989484820 /* phi / golden ratio */ #endif +#ifndef M_PI +#define M_PI 3.14159265358979323846 /* pi */ +#endif +#ifndef M_PI_2 +#define M_PI_2 1.57079632679489661923 /* pi/2 */ +#endif +#ifndef M_SQRT1_2 +#define M_SQRT1_2 0.70710678118654752440 /* 1/sqrt(2) */ +#endif +#ifndef M_SQRT2 +#define M_SQRT2 1.41421356237309504880 /* sqrt(2) */ +#endif #ifndef NAN #define NAN av_int2float(0x7fc00000) #endif @@ -52,12 +73,14 @@ enum AVRounding { AV_ROUND_DOWN = 2, ///< Round toward -infinity. AV_ROUND_UP = 3, ///< Round toward +infinity. AV_ROUND_NEAR_INF = 5, ///< Round to nearest and halfway cases away from zero. + AV_ROUND_PASS_MINMAX = 8192, ///< Flag to pass INT64_MIN/MAX through instead of rescaling, this avoids special cases for AV_NOPTS_VALUE }; /** - * Return the greatest common divisor of a and b. - * If both a and b are 0 or either or both are <0 then behavior is - * undefined. + * Compute the greatest common divisor of a and b. + * + * @return gcd of a and b up to sign; if a >= 0 and b >= 0, return value is >= 0; + * if a == 0 and b == 0, returns 0. */ int64_t av_const av_gcd(int64_t a, int64_t b); @@ -70,6 +93,9 @@ int64_t av_rescale(int64_t a, int64_t b, int64_t c) av_const; /** * Rescale a 64-bit integer with specified rounding. * A simple a*b/c isn't possible as it can overflow. + * + * @return rescaled value a, or if AV_ROUND_PASS_MINMAX is set and a is + * INT64_MIN or INT64_MAX then a is passed through unchanged. */ int64_t av_rescale_rnd(int64_t a, int64_t b, int64_t c, enum AVRounding) av_const; @@ -80,6 +106,9 @@ int64_t av_rescale_q(int64_t a, AVRational bq, AVRational cq) av_const; /** * Rescale a 64-bit integer by 2 rational numbers with specified rounding. + * + * @return rescaled value a, or if AV_ROUND_PASS_MINMAX is set and a is + * INT64_MIN or INT64_MAX then a is passed through unchanged. */ int64_t av_rescale_q_rnd(int64_t a, AVRational bq, AVRational cq, enum AVRounding) av_const; @@ -105,6 +134,31 @@ int av_compare_ts(int64_t ts_a, AVRational tb_a, int64_t ts_b, AVRational tb_b); int64_t av_compare_mod(uint64_t a, uint64_t b, uint64_t mod); /** + * Rescale a timestamp while preserving known durations. + * + * @param in_ts Input timestamp + * @param in_tb Input timebase + * @param fs_tb Duration and *last timebase + * @param duration duration till the next call + * @param out_tb Output timebase + */ +int64_t av_rescale_delta(AVRational in_tb, int64_t in_ts, AVRational fs_tb, int duration, int64_t *last, AVRational out_tb); + +/** + * Add a value to a timestamp. + * + * This function guarantees that when the same value is repeatly added that + * no accumulation of rounding errors occurs. + * + * @param ts Input timestamp + * @param ts_tb Input timestamp timebase + * @param inc value to add to ts + * @param inc_tb inc timebase + */ +int64_t av_add_stable(AVRational ts_tb, int64_t ts, AVRational inc_tb, int64_t inc); + + + /** * @} */ diff --git a/libavutil/md5-test.c b/libavutil/md5-test.c index 8d5dd1e104..547f74b880 100644 --- a/libavutil/md5-test.c +++ b/libavutil/md5-test.c @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -33,7 +33,8 @@ int main(void) { uint8_t md5val[16]; int i; - uint8_t in[1000]; + volatile uint8_t in[1000]; // volatile to workaround http://llvm.org/bugs/show_bug.cgi?id=20849 + // FIXME remove volatile once it has been fixed and all fate clients are updated for (i = 0; i < 1000; i++) in[i] = i * i; diff --git a/libavutil/md5.c b/libavutil/md5.c index e02a9ba0bf..057085b937 100644 --- a/libavutil/md5.c +++ b/libavutil/md5.c @@ -13,20 +13,20 @@ * If you use gcc, then version 4.1 or later and -fomit-frame-pointer is * strongly recommended. * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -43,6 +43,8 @@ typedef struct AVMD5 { uint32_t ABCD[4]; } AVMD5; +const int av_md5_size = sizeof(AVMD5); + struct AVMD5 *av_md5_alloc(void) { return av_mallocz(sizeof(struct AVMD5)); @@ -84,56 +86,62 @@ static const uint32_t T[64] = { // T[i]= fabs(sin(i+1)<<32) \ if (i < 32) { \ if (i < 16) \ - a += (d ^ (b & (c ^ d))) + X[i & 15]; \ + a += (d ^ (b & (c ^ d))) + X[ i & 15]; \ else \ - a += (c ^ (d & (c ^ b))) + X[(1 + 5 * i) & 15]; \ + a += ((d & b) | (~d & c)) + X[(1 + 5*i) & 15]; \ } else { \ if (i < 48) \ - a += (b ^ c ^ d) + X[(5 + 3 * i) & 15]; \ + a += (b ^ c ^ d) + X[(5 + 3*i) & 15]; \ else \ - a += (c ^ (b | ~d)) + X[(7 * i) & 15]; \ + a += (c ^ (b | ~d)) + X[( 7*i) & 15]; \ } \ a = b + (a << t | a >> (32 - t)); \ } while (0) -static void body(uint32_t ABCD[4], uint32_t X[16]) +static void body(uint32_t ABCD[4], uint32_t *src, int nblocks) { - int t; int i av_unused; - unsigned int a = ABCD[3]; - unsigned int b = ABCD[2]; - unsigned int c = ABCD[1]; - unsigned int d = ABCD[0]; + int n; + uint32_t a, b, c, d, t, *X; + + for (n = 0; n < nblocks; n++) { + a = ABCD[3]; + b = ABCD[2]; + c = ABCD[1]; + d = ABCD[0]; + + X = src + n * 16; #if HAVE_BIGENDIAN - for (i = 0; i < 16; i++) - X[i] = av_bswap32(X[i]); + for (i = 0; i < 16; i++) + X[i] = av_bswap32(X[i]); #endif #if CONFIG_SMALL - for (i = 0; i < 64; i++) { - CORE(i, a, b, c, d); - t = d; - d = c; - c = b; - b = a; - a = t; - } + for (i = 0; i < 64; i++) { + CORE(i, a, b, c, d); + t = d; + d = c; + c = b; + b = a; + a = t; + } #else #define CORE2(i) \ - CORE(i, a, b, c, d); CORE((i + 1), d, a, b, c); \ - CORE((i + 2), c, d, a, b); CORE((i + 3), b, c, d, a) + CORE(i, a, b, c, d); CORE((i + 1), d, a, b, c); \ + CORE((i + 2), c, d, a, b); CORE((i + 3), b, c, d, a) #define CORE4(i) CORE2(i); CORE2((i + 4)); CORE2((i + 8)); CORE2((i + 12)) - CORE4(0); - CORE4(16); - CORE4(32); - CORE4(48); + CORE4(0); + CORE4(16); + CORE4(32); + CORE4(48); #endif - ABCD[0] += d; - ABCD[1] += c; - ABCD[2] += b; - ABCD[3] += a; + ABCD[0] += d; + ABCD[1] += c; + ABCD[2] += b; + ABCD[3] += a; + } } void av_md5_init(AVMD5 *ctx) @@ -146,20 +154,39 @@ void av_md5_init(AVMD5 *ctx) ctx->ABCD[3] = 0x67452301; } -void av_md5_update(AVMD5 *ctx, const uint8_t *src, const int len) +void av_md5_update(AVMD5 *ctx, const uint8_t *src, int len) { - int i, j; + const uint8_t *end; + int j; j = ctx->len & 63; ctx->len += len; - for (i = 0; i < len; i++) { - ctx->block[j++] = src[i]; - if (j == 64) { - body(ctx->ABCD, (uint32_t *) ctx->block); - j = 0; + if (j) { + int cnt = FFMIN(len, 64 - j); + memcpy(ctx->block + j, src, cnt); + src += cnt; + len -= cnt; + if (j + cnt < 64) + return; + body(ctx->ABCD, (uint32_t *)ctx->block, 1); + } + + end = src + (len & ~63); + if (HAVE_BIGENDIAN || (!HAVE_FAST_UNALIGNED && ((intptr_t)src & 3))) { + while (src < end) { + memcpy(ctx->block, src, 64); + body(ctx->ABCD, (uint32_t *) ctx->block, 1); + src += 64; } + } else { + int nblocks = len / 64; + body(ctx->ABCD, (uint32_t *)src, nblocks); + src = end; } + len &= 63; + if (len > 0) + memcpy(ctx->block, src, len); } void av_md5_final(AVMD5 *ctx, uint8_t *dst) diff --git a/libavutil/md5.h b/libavutil/md5.h index c26318c01d..79702c88c2 100644 --- a/libavutil/md5.h +++ b/libavutil/md5.h @@ -1,20 +1,20 @@ /* * copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -32,12 +32,46 @@ * @{ */ +extern const int av_md5_size; + struct AVMD5; +/** + * Allocate an AVMD5 context. + */ struct AVMD5 *av_md5_alloc(void); + +/** + * Initialize MD5 hashing. + * + * @param ctx pointer to the function context (of size av_md5_size) + */ void av_md5_init(struct AVMD5 *ctx); -void av_md5_update(struct AVMD5 *ctx, const uint8_t *src, const int len); + +/** + * Update hash value. + * + * @param ctx hash function context + * @param src input data to update hash with + * @param len input data length + */ +void av_md5_update(struct AVMD5 *ctx, const uint8_t *src, int len); + +/** + * Finish hashing and output digest value. + * + * @param ctx hash function context + * @param dst buffer where output digest value is stored + */ void av_md5_final(struct AVMD5 *ctx, uint8_t *dst); + +/** + * Hash an array of data. + * + * @param dst The output buffer to write the digest into + * @param src The data to hash + * @param len The length of the data, in bytes + */ void av_md5_sum(uint8_t *dst, const uint8_t *src, const int len); /** diff --git a/libavutil/mem.c b/libavutil/mem.c index 15c28808c1..8dfaad8271 100644 --- a/libavutil/mem.c +++ b/libavutil/mem.c @@ -2,20 +2,20 @@ * default memory allocator for libavutil * Copyright (c) 2002 Fabrice Bellard * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -24,6 +24,8 @@ * default memory allocator for libavutil */ +#define _XOPEN_SOURCE 600 + #include "config.h" #include <limits.h> @@ -34,8 +36,10 @@ #include <malloc.h> #endif +#include "avassert.h" #include "avutil.h" #include "common.h" +#include "dynarray.h" #include "intreadwrite.h" #include "mem.h" @@ -55,9 +59,20 @@ void free(void *ptr); #endif /* MALLOC_PREFIX */ -/* You can redefine av_malloc and av_free in your project to use your - * memory allocator. You do not need to suppress this file because the - * linker will do it automatically. */ +#include "mem_internal.h" + +#define ALIGN (HAVE_AVX ? 32 : 16) + +/* NOTE: if you want to override these functions with your own + * implementations (not recommended) you have to link libav* as + * dynamic libraries and remove -Wl,-Bsymbolic from the linker flags. + * Note that this will cost performance. */ + +static size_t max_alloc_size= INT_MAX; + +void av_max_alloc(size_t max){ + max_alloc_size = max; +} void *av_malloc(size_t size) { @@ -67,23 +82,28 @@ void *av_malloc(size_t size) #endif /* let's disallow possibly ambiguous cases */ - if (size > (INT_MAX - 32) || !size) + if (size > (max_alloc_size - 32)) return NULL; #if CONFIG_MEMALIGN_HACK - ptr = malloc(size + 32); + ptr = malloc(size + ALIGN); if (!ptr) return ptr; - diff = ((-(long)ptr - 1) & 31) + 1; + diff = ((~(long)ptr)&(ALIGN - 1)) + 1; ptr = (char *)ptr + diff; ((char *)ptr)[-1] = diff; #elif HAVE_POSIX_MEMALIGN - if (posix_memalign(&ptr, 32, size)) + if (size) //OS X on SDK 10.6 has a broken posix_memalign implementation + if (posix_memalign(&ptr, ALIGN, size)) ptr = NULL; #elif HAVE_ALIGNED_MALLOC - ptr = _aligned_malloc(size, 32); + ptr = _aligned_malloc(size, ALIGN); #elif HAVE_MEMALIGN - ptr = memalign(32, size); +#ifndef __DJGPP__ + ptr = memalign(ALIGN, size); +#else + ptr = memalign(size, ALIGN); +#endif /* Why 64? * Indeed, we should align it: * on 4 for 386 @@ -111,6 +131,14 @@ void *av_malloc(size_t size) #else ptr = malloc(size); #endif + if(!ptr && !size) { + size = 1; + ptr= av_malloc(1); + } +#if CONFIG_MEMORY_POISONING + if (ptr) + memset(ptr, FF_MEMORY_POISON, size); +#endif return ptr; } @@ -121,7 +149,7 @@ void *av_realloc(void *ptr, size_t size) #endif /* let's disallow possibly ambiguous cases */ - if (size > (INT_MAX - 16)) + if (size > (max_alloc_size - 32)) return NULL; #if CONFIG_MEMALIGN_HACK @@ -129,14 +157,33 @@ void *av_realloc(void *ptr, size_t size) if (!ptr) return av_malloc(size); diff = ((char *)ptr)[-1]; - return (char *)realloc((char *)ptr - diff, size + diff) + diff; + av_assert0(diff>0 && diff<=ALIGN); + ptr = realloc((char *)ptr - diff, size + diff); + if (ptr) + ptr = (char *)ptr + diff; + return ptr; #elif HAVE_ALIGNED_MALLOC - return _aligned_realloc(ptr, size, 32); + return _aligned_realloc(ptr, size + !size, ALIGN); #else - return realloc(ptr, size); + return realloc(ptr, size + !size); #endif } +void *av_realloc_f(void *ptr, size_t nelem, size_t elsize) +{ + size_t size; + void *r; + + if (av_size_mult(elsize, nelem, &size)) { + av_free(ptr); + return NULL; + } + r = av_realloc(ptr, size); + if (!r && size) + av_free(ptr); + return r; +} + int av_reallocp(void *ptr, size_t size) { void *val; @@ -169,29 +216,23 @@ int av_reallocp_array(void *ptr, size_t nmemb, size_t size) { void *val; - if (!size || nmemb >= INT_MAX / size) - return AVERROR(ENOMEM); - if (!nmemb) { - av_freep(ptr); - return 0; - } - memcpy(&val, ptr, sizeof(val)); - val = av_realloc(val, nmemb * size); - if (!val) { - av_freep(ptr); + val = av_realloc_f(val, nmemb, size); + memcpy(ptr, &val, sizeof(val)); + if (!val && nmemb && size) return AVERROR(ENOMEM); - } - memcpy(ptr, &val, sizeof(val)); return 0; } void av_free(void *ptr) { #if CONFIG_MEMALIGN_HACK - if (ptr) - free((char *)ptr - ((char *)ptr)[-1]); + if (ptr) { + int v= ((char *)ptr)[-1]; + av_assert0(v>0 && v<=ALIGN); + free((char *)ptr - v); + } #elif HAVE_ALIGNED_MALLOC _aligned_free(ptr); #else @@ -216,11 +257,18 @@ void *av_mallocz(size_t size) return ptr; } +void *av_calloc(size_t nmemb, size_t size) +{ + if (size <= 0 || nmemb >= INT_MAX / size) + return NULL; + return av_mallocz(nmemb * size); +} + char *av_strdup(const char *s) { char *ptr = NULL; if (s) { - int len = strlen(s) + 1; + size_t len = strlen(s) + 1; ptr = av_realloc(NULL, len); if (ptr) memcpy(ptr, s, len); @@ -248,6 +296,63 @@ char *av_strndup(const char *s, size_t len) return ret; } +void *av_memdup(const void *p, size_t size) +{ + void *ptr = NULL; + if (p) { + ptr = av_malloc(size); + if (ptr) + memcpy(ptr, p, size); + } + return ptr; +} + +int av_dynarray_add_nofree(void *tab_ptr, int *nb_ptr, void *elem) +{ + void **tab; + memcpy(&tab, tab_ptr, sizeof(tab)); + + AV_DYNARRAY_ADD(INT_MAX, sizeof(*tab), tab, *nb_ptr, { + tab[*nb_ptr] = elem; + memcpy(tab_ptr, &tab, sizeof(tab)); + }, { + return AVERROR(ENOMEM); + }); + return 0; +} + +void av_dynarray_add(void *tab_ptr, int *nb_ptr, void *elem) +{ + void **tab; + memcpy(&tab, tab_ptr, sizeof(tab)); + + AV_DYNARRAY_ADD(INT_MAX, sizeof(*tab), tab, *nb_ptr, { + tab[*nb_ptr] = elem; + memcpy(tab_ptr, &tab, sizeof(tab)); + }, { + *nb_ptr = 0; + av_freep(tab_ptr); + }); +} + +void *av_dynarray2_add(void **tab_ptr, int *nb_ptr, size_t elem_size, + const uint8_t *elem_data) +{ + uint8_t *tab_elem_data = NULL; + + AV_DYNARRAY_ADD(INT_MAX, elem_size, *tab_ptr, *nb_ptr, { + tab_elem_data = (uint8_t *)*tab_ptr + (*nb_ptr) * elem_size; + if (elem_data) + memcpy(tab_elem_data, elem_data, elem_size); + else if (CONFIG_MEMORY_POISONING) + memset(tab_elem_data, FF_MEMORY_POISON, elem_size); + }, { + av_freep(tab_ptr); + *nb_ptr = 0; + }); + return tab_elem_data; +} + static void fill16(uint8_t *dst, int len) { uint32_t v = AV_RN16(dst - 2); @@ -377,7 +482,7 @@ void *av_fast_realloc(void *ptr, unsigned int *size, size_t min_size) if (min_size < *size) return ptr; - min_size = FFMAX(17 * min_size / 16 + 32, min_size); + min_size = FFMAX(min_size + min_size / 16 + 32, min_size); ptr = av_realloc(ptr, min_size); /* we could set this to the unmodified min_size but this is safer @@ -393,13 +498,10 @@ void *av_fast_realloc(void *ptr, unsigned int *size, size_t min_size) void av_fast_malloc(void *ptr, unsigned int *size, size_t min_size) { - void **p = ptr; - if (min_size < *size) - return; - min_size = FFMAX(17 * min_size / 16 + 32, min_size); - av_free(*p); - *p = av_malloc(min_size); - if (!*p) - min_size = 0; - *size = min_size; + ff_fast_malloc(ptr, size, min_size, 0); +} + +void av_fast_mallocz(void *ptr, unsigned int *size, size_t min_size) +{ + ff_fast_malloc(ptr, size, min_size, 1); } diff --git a/libavutil/mem.h b/libavutil/mem.h index 9f667c270b..d25b3229b7 100644 --- a/libavutil/mem.h +++ b/libavutil/mem.h @@ -1,20 +1,20 @@ /* * copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -30,6 +30,7 @@ #include <stdint.h> #include "attributes.h" +#include "error.h" #include "avutil.h" /** @@ -38,7 +39,7 @@ */ -#if defined(__ICC) && __ICC < 1200 || defined(__SUNPRO_C) +#if defined(__INTEL_COMPILER) && __INTEL_COMPILER < 1110 || defined(__SUNPRO_C) #define DECLARE_ALIGNED(n,t,v) t __attribute__ ((aligned (n))) v #define DECLARE_ASM_CONST(n,t,v) const t __attribute__ ((aligned (n))) v #elif defined(__TI_COMPILER_VERSION__) @@ -118,6 +119,16 @@ void *av_realloc(void *ptr, size_t size) av_alloc_size(2); /** * Allocate or reallocate a block of memory. + * This function does the same thing as av_realloc, except: + * - It takes two arguments and checks the result of the multiplication for + * integer overflow. + * - It frees the input block in case of failure, thus avoiding the memory + * leak with the classic "buf = realloc(buf); if (!buf) return -1;". + */ +void *av_realloc_f(void *ptr, size_t nelem, size_t elsize); + +/** + * Allocate or reallocate a block of memory. * If *ptr is NULL and size > 0, allocate a new block. If * size is zero, free the memory block pointed to by ptr. * @param ptr Pointer to a pointer to a memory block already allocated @@ -133,6 +144,7 @@ void *av_realloc(void *ptr, size_t size) av_alloc_size(2); * The situation is undefined according to POSIX and may crash with * some libc implementations. */ +av_warn_unused_result int av_reallocp(void *ptr, size_t size); /** @@ -194,6 +206,18 @@ void av_free(void *ptr); void *av_mallocz(size_t size) av_malloc_attrib av_alloc_size(1); /** + * Allocate a block of nmemb * size bytes with alignment suitable for all + * memory accesses (including vectors if available on the CPU) and + * zero all the bytes of the block. + * The allocation will fail if nmemb * size is greater than or equal + * to INT_MAX. + * @param nmemb + * @param size + * @return Pointer to the allocated block, NULL if it cannot be allocated. + */ +void *av_calloc(size_t nmemb, size_t size) av_malloc_attrib; + +/** * Allocate a block of size * nmemb bytes with av_mallocz(). * @param nmemb Number of elements * @param size Size of the single element @@ -228,18 +252,110 @@ char *av_strdup(const char *s) av_malloc_attrib; char *av_strndup(const char *s, size_t len) av_malloc_attrib; /** + * Duplicate the buffer p. + * @param p buffer to be duplicated + * @return Pointer to a newly allocated buffer containing a + * copy of p or NULL if the buffer cannot be allocated. + */ +void *av_memdup(const void *p, size_t size); + +/** * Free a memory block which has been allocated with av_malloc(z)() or * av_realloc() and set the pointer pointing to it to NULL. * @param ptr Pointer to the pointer to the memory block which should * be freed. + * @note passing a pointer to a NULL pointer is safe and leads to no action. * @see av_free() */ void av_freep(void *ptr); /** + * Add an element to a dynamic array. + * + * The array to grow is supposed to be an array of pointers to + * structures, and the element to add must be a pointer to an already + * allocated structure. + * + * The array is reallocated when its size reaches powers of 2. + * Therefore, the amortized cost of adding an element is constant. + * + * In case of success, the pointer to the array is updated in order to + * point to the new grown array, and the number pointed to by nb_ptr + * is incremented. + * In case of failure, the array is freed, *tab_ptr is set to NULL and + * *nb_ptr is set to 0. + * + * @param tab_ptr pointer to the array to grow + * @param nb_ptr pointer to the number of elements in the array + * @param elem element to add + * @see av_dynarray_add_nofree(), av_dynarray2_add() + */ +void av_dynarray_add(void *tab_ptr, int *nb_ptr, void *elem); + +/** + * Add an element to a dynamic array. + * + * Function has the same functionality as av_dynarray_add(), + * but it doesn't free memory on fails. It returns error code + * instead and leave current buffer untouched. + * + * @param tab_ptr pointer to the array to grow + * @param nb_ptr pointer to the number of elements in the array + * @param elem element to add + * @return >=0 on success, negative otherwise. + * @see av_dynarray_add(), av_dynarray2_add() + */ +av_warn_unused_result +int av_dynarray_add_nofree(void *tab_ptr, int *nb_ptr, void *elem); + +/** + * Add an element of size elem_size to a dynamic array. + * + * The array is reallocated when its number of elements reaches powers of 2. + * Therefore, the amortized cost of adding an element is constant. + * + * In case of success, the pointer to the array is updated in order to + * point to the new grown array, and the number pointed to by nb_ptr + * is incremented. + * In case of failure, the array is freed, *tab_ptr is set to NULL and + * *nb_ptr is set to 0. + * + * @param tab_ptr pointer to the array to grow + * @param nb_ptr pointer to the number of elements in the array + * @param elem_size size in bytes of the elements in the array + * @param elem_data pointer to the data of the element to add. If NULL, the space of + * the new added element is not filled. + * @return pointer to the data of the element to copy in the new allocated space. + * If NULL, the new allocated space is left uninitialized." + * @see av_dynarray_add(), av_dynarray_add_nofree() + */ +void *av_dynarray2_add(void **tab_ptr, int *nb_ptr, size_t elem_size, + const uint8_t *elem_data); + +/** + * Multiply two size_t values checking for overflow. + * @return 0 if success, AVERROR(EINVAL) if overflow. + */ +static inline int av_size_mult(size_t a, size_t b, size_t *r) +{ + size_t t = a * b; + /* Hack inspired from glibc: only try the division if nelem and elsize + * are both greater than sqrt(SIZE_MAX). */ + if ((a | b) >= ((size_t)1 << (sizeof(size_t) * 4)) && a && t / a != b) + return AVERROR(EINVAL); + *r = t; + return 0; +} + +/** + * Set the maximum size that may me allocated in one block. + */ +void av_max_alloc(size_t max); + +/** * deliberately overlapping memcpy implementation * @param dst destination buffer - * @param back how many bytes back we start (the initial size of the overlapping window) + * @param back how many bytes back we start (the initial size of the overlapping window), must be > 0 * @param cnt number of bytes to copy, must be >= 0 * * cnt > back is valid, this will copy the bytes we just copied, @@ -269,6 +385,21 @@ void *av_fast_realloc(void *ptr, unsigned int *size, size_t min_size); void av_fast_malloc(void *ptr, unsigned int *size, size_t min_size); /** + * Allocate a buffer, reusing the given one if large enough. + * + * All newly allocated space is initially cleared + * Contrary to av_fast_realloc the current buffer contents might not be + * preserved and on error the old buffer is freed, thus no special + * handling to avoid memleaks is necessary. + * + * @param ptr pointer to pointer to already allocated buffer, overwritten with pointer to new buffer + * @param size size of the buffer *ptr points to + * @param min_size minimum size of *ptr buffer after returning, *ptr will be NULL and + * *size 0 if an error occurred. + */ +void av_fast_mallocz(void *ptr, unsigned int *size, size_t min_size); + +/** * @} */ diff --git a/libavutil/mem_internal.h b/libavutil/mem_internal.h new file mode 100644 index 0000000000..6fdbcb016e --- /dev/null +++ b/libavutil/mem_internal.h @@ -0,0 +1,45 @@ +/* + * Copyright (c) 2002 Fabrice Bellard + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#ifndef AVUTIL_MEM_INTERNAL_H +#define AVUTIL_MEM_INTERNAL_H + +#include "avassert.h" +#include "mem.h" + +static inline int ff_fast_malloc(void *ptr, unsigned int *size, size_t min_size, int zero_realloc) +{ + void *val; + + memcpy(&val, ptr, sizeof(val)); + if (min_size <= *size) { + av_assert0(val || !min_size); + return 0; + } + min_size = FFMAX(min_size + min_size / 16 + 32, min_size); + av_freep(ptr); + val = zero_realloc ? av_mallocz(min_size) : av_malloc(min_size); + memcpy(ptr, &val, sizeof(val)); + if (!val) + min_size = 0; + *size = min_size; + return 1; +} +#endif /* AVUTIL_MEM_INTERNAL_H */ diff --git a/libavutil/mips/Makefile b/libavutil/mips/Makefile new file mode 100644 index 0000000000..dbfa5aa341 --- /dev/null +++ b/libavutil/mips/Makefile @@ -0,0 +1 @@ +OBJS += mips/float_dsp_mips.o diff --git a/libavutil/mips/asmdefs.h b/libavutil/mips/asmdefs.h new file mode 100644 index 0000000000..fdf82a0107 --- /dev/null +++ b/libavutil/mips/asmdefs.h @@ -0,0 +1,46 @@ +/* + * Copyright (c) 2015 Imagination Technologies Ltd + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +/** + * @file + * MIPS assembly defines from sys/asm.h but rewritten for use with C inline + * assembly (rather than from within .s files). + */ + +#ifndef AVUTIL_MIPS_ASMDEFS_H +#define AVUTIL_MIPS_ASMDEFS_H + +#if defined(_ABI64) && _MIPS_SIM == _ABI64 +# define PTRSIZE " 8 " +# define PTRLOG " 3 " +# define PTR_ADDU "daddu " +# define PTR_ADDIU "daddiu " +# define PTR_SUBU "dsubu " +# define PTR_L "ld " +#else +# define PTRSIZE " 4 " +# define PTRLOG " 2 " +# define PTR_ADDU "addu " +# define PTR_ADDIU "addiu " +# define PTR_SUBU "subu " +# define PTR_L "lw " +#endif + +#endif /* AVCODEC_MIPS_ASMDEFS_H */ diff --git a/libavutil/mips/float_dsp_mips.c b/libavutil/mips/float_dsp_mips.c new file mode 100644 index 0000000000..0943d6f343 --- /dev/null +++ b/libavutil/mips/float_dsp_mips.c @@ -0,0 +1,356 @@ +/* + * Copyright (c) 2012 + * MIPS Technologies, Inc., California. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * 3. Neither the name of the MIPS Technologies, Inc., nor the names of its + * contributors may be used to endorse or promote products derived from + * this software without specific prior written permission. + * + * THIS SOFTWARE IS PROVIDED BY THE MIPS TECHNOLOGIES, INC. ``AS IS'' AND + * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE + * ARE DISCLAIMED. IN NO EVENT SHALL THE MIPS TECHNOLOGIES, INC. BE LIABLE + * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT + * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY + * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF + * SUCH DAMAGE. + * + * Author: Branimir Vasic (bvasic@mips.com) + * Author: Zoran Lukic (zoranl@mips.com) + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +/** + * @file + * Reference: libavutil/float_dsp.c + */ + +#include "config.h" +#include "libavutil/float_dsp.h" +#include "libavutil/mips/asmdefs.h" + +#if HAVE_INLINE_ASM && HAVE_MIPSFPU +#if !HAVE_MIPS32R6 && !HAVE_MIPS64R6 +static void vector_fmul_mips(float *dst, const float *src0, const float *src1, + int len) +{ + int i; + + if (len & 3) { + for (i = 0; i < len; i++) + dst[i] = src0[i] * src1[i]; + } else { + float *d = (float *)dst; + float *d_end = d + len; + float *s0 = (float *)src0; + float *s1 = (float *)src1; + + float src0_0, src0_1, src0_2, src0_3; + float src1_0, src1_1, src1_2, src1_3; + + __asm__ volatile ( + "1: \n\t" + "lwc1 %[src0_0], 0(%[s0]) \n\t" + "lwc1 %[src1_0], 0(%[s1]) \n\t" + "lwc1 %[src0_1], 4(%[s0]) \n\t" + "lwc1 %[src1_1], 4(%[s1]) \n\t" + "lwc1 %[src0_2], 8(%[s0]) \n\t" + "lwc1 %[src1_2], 8(%[s1]) \n\t" + "lwc1 %[src0_3], 12(%[s0]) \n\t" + "lwc1 %[src1_3], 12(%[s1]) \n\t" + "mul.s %[src0_0], %[src0_0], %[src1_0] \n\t" + "mul.s %[src0_1], %[src0_1], %[src1_1] \n\t" + "mul.s %[src0_2], %[src0_2], %[src1_2] \n\t" + "mul.s %[src0_3], %[src0_3], %[src1_3] \n\t" + "swc1 %[src0_0], 0(%[d]) \n\t" + "swc1 %[src0_1], 4(%[d]) \n\t" + "swc1 %[src0_2], 8(%[d]) \n\t" + "swc1 %[src0_3], 12(%[d]) \n\t" + PTR_ADDIU "%[s0], %[s0], 16 \n\t" + PTR_ADDIU "%[s1], %[s1], 16 \n\t" + PTR_ADDIU "%[d], %[d], 16 \n\t" + "bne %[d], %[d_end], 1b \n\t" + + : [src0_0]"=&f"(src0_0), [src0_1]"=&f"(src0_1), + [src0_2]"=&f"(src0_2), [src0_3]"=&f"(src0_3), + [src1_0]"=&f"(src1_0), [src1_1]"=&f"(src1_1), + [src1_2]"=&f"(src1_2), [src1_3]"=&f"(src1_3), + [d]"+r"(d), [s0]"+r"(s0), [s1]"+r"(s1) + : [d_end]"r"(d_end) + : "memory" + ); + } +} + +static void vector_fmul_scalar_mips(float *dst, const float *src, float mul, + int len) +{ + float temp0, temp1, temp2, temp3; + float *local_src = (float*)src; + float *end = local_src + len; + + /* loop unrolled 4 times */ + __asm__ volatile( + ".set push \n\t" + ".set noreorder \n\t" + "1: \n\t" + "lwc1 %[temp0], 0(%[src]) \n\t" + "lwc1 %[temp1], 4(%[src]) \n\t" + "lwc1 %[temp2], 8(%[src]) \n\t" + "lwc1 %[temp3], 12(%[src]) \n\t" + PTR_ADDIU "%[dst], %[dst], 16 \n\t" + "mul.s %[temp0], %[temp0], %[mul] \n\t" + "mul.s %[temp1], %[temp1], %[mul] \n\t" + "mul.s %[temp2], %[temp2], %[mul] \n\t" + "mul.s %[temp3], %[temp3], %[mul] \n\t" + PTR_ADDIU "%[src], %[src], 16 \n\t" + "swc1 %[temp0], -16(%[dst]) \n\t" + "swc1 %[temp1], -12(%[dst]) \n\t" + "swc1 %[temp2], -8(%[dst]) \n\t" + "bne %[src], %[end], 1b \n\t" + " swc1 %[temp3], -4(%[dst]) \n\t" + ".set pop \n\t" + + : [temp0]"=&f"(temp0), [temp1]"=&f"(temp1), + [temp2]"=&f"(temp2), [temp3]"=&f"(temp3), + [dst]"+r"(dst), [src]"+r"(local_src) + : [end]"r"(end), [mul]"f"(mul) + : "memory" + ); +} + +static void vector_fmul_window_mips(float *dst, const float *src0, + const float *src1, const float *win, int len) +{ + float * dst_j, *win_j, *src0_i, *src1_j, *dst_i, *win_i; + float temp, temp1, temp2, temp3; + float s0, s01, s1, s11; + float wi, wi1, wi2, wi3; + float wj, wj1, wj2, wj3; + const float * lp_end = win + len; + + win_i = (float *)win; + win_j = (float *)(win + 2 * len -1); + src1_j = (float *)(src1 + len - 1); + src0_i = (float *)src0; + dst_i = (float *)dst; + dst_j = (float *)(dst + 2 * len -1); + + /* loop unrolled 4 times */ + __asm__ volatile ( + "1:" + "lwc1 %[s1], 0(%[src1_j]) \n\t" + "lwc1 %[wi], 0(%[win_i]) \n\t" + "lwc1 %[wj], 0(%[win_j]) \n\t" + "lwc1 %[s11], -4(%[src1_j]) \n\t" + "lwc1 %[wi1], 4(%[win_i]) \n\t" + "lwc1 %[wj1], -4(%[win_j]) \n\t" + "lwc1 %[s0], 0(%[src0_i]) \n\t" + "lwc1 %[s01], 4(%[src0_i]) \n\t" + "mul.s %[temp], %[s1], %[wi] \n\t" + "mul.s %[temp1], %[s1], %[wj] \n\t" + "mul.s %[temp2], %[s11], %[wi1] \n\t" + "mul.s %[temp3], %[s11], %[wj1] \n\t" + "lwc1 %[s1], -8(%[src1_j]) \n\t" + "lwc1 %[wi2], 8(%[win_i]) \n\t" + "lwc1 %[wj2], -8(%[win_j]) \n\t" + "lwc1 %[s11], -12(%[src1_j]) \n\t" + "msub.s %[temp], %[temp], %[s0], %[wj] \n\t" + "madd.s %[temp1], %[temp1], %[s0], %[wi] \n\t" + "msub.s %[temp2], %[temp2], %[s01], %[wj1] \n\t" + "madd.s %[temp3], %[temp3], %[s01], %[wi1] \n\t" + "lwc1 %[wi3], 12(%[win_i]) \n\t" + "lwc1 %[wj3], -12(%[win_j]) \n\t" + "lwc1 %[s0], 8(%[src0_i]) \n\t" + "lwc1 %[s01], 12(%[src0_i]) \n\t" + PTR_ADDIU "%[src1_j],-16 \n\t" + PTR_ADDIU "%[win_i],16 \n\t" + PTR_ADDIU "%[win_j],-16 \n\t" + PTR_ADDIU "%[src0_i],16 \n\t" + "swc1 %[temp], 0(%[dst_i]) \n\t" /* dst[i] = s0*wj - s1*wi; */ + "swc1 %[temp1], 0(%[dst_j]) \n\t" /* dst[j] = s0*wi + s1*wj; */ + "swc1 %[temp2], 4(%[dst_i]) \n\t" /* dst[i+1] = s01*wj1 - s11*wi1; */ + "swc1 %[temp3], -4(%[dst_j]) \n\t" /* dst[j-1] = s01*wi1 + s11*wj1; */ + "mul.s %[temp], %[s1], %[wi2] \n\t" + "mul.s %[temp1], %[s1], %[wj2] \n\t" + "mul.s %[temp2], %[s11], %[wi3] \n\t" + "mul.s %[temp3], %[s11], %[wj3] \n\t" + "msub.s %[temp], %[temp], %[s0], %[wj2] \n\t" + "madd.s %[temp1], %[temp1], %[s0], %[wi2] \n\t" + "msub.s %[temp2], %[temp2], %[s01], %[wj3] \n\t" + "madd.s %[temp3], %[temp3], %[s01], %[wi3] \n\t" + "swc1 %[temp], 8(%[dst_i]) \n\t" /* dst[i+2] = s0*wj2 - s1*wi2; */ + "swc1 %[temp1], -8(%[dst_j]) \n\t" /* dst[j-2] = s0*wi2 + s1*wj2; */ + "swc1 %[temp2], 12(%[dst_i]) \n\t" /* dst[i+2] = s01*wj3 - s11*wi3; */ + "swc1 %[temp3], -12(%[dst_j]) \n\t" /* dst[j-3] = s01*wi3 + s11*wj3; */ + PTR_ADDIU "%[dst_i],16 \n\t" + PTR_ADDIU "%[dst_j],-16 \n\t" + "bne %[win_i], %[lp_end], 1b \n\t" + : [temp]"=&f"(temp), [temp1]"=&f"(temp1), [temp2]"=&f"(temp2), + [temp3]"=&f"(temp3), [src0_i]"+r"(src0_i), [win_i]"+r"(win_i), + [src1_j]"+r"(src1_j), [win_j]"+r"(win_j), [dst_i]"+r"(dst_i), + [dst_j]"+r"(dst_j), [s0] "=&f"(s0), [s01]"=&f"(s01), [s1] "=&f"(s1), + [s11]"=&f"(s11), [wi] "=&f"(wi), [wj] "=&f"(wj), [wi2]"=&f"(wi2), + [wj2]"=&f"(wj2), [wi3]"=&f"(wi3), [wj3]"=&f"(wj3), [wi1]"=&f"(wi1), + [wj1]"=&f"(wj1) + : [lp_end]"r"(lp_end) + : "memory" + ); +} + +static void butterflies_float_mips(float *av_restrict v1, float *av_restrict v2, + int len) +{ + float temp0, temp1, temp2, temp3, temp4; + float temp5, temp6, temp7, temp8, temp9; + float temp10, temp11, temp12, temp13, temp14, temp15; + int pom; + pom = (len >> 2)-1; + + /* loop unrolled 4 times */ + __asm__ volatile ( + "lwc1 %[temp0], 0(%[v1]) \n\t" + "lwc1 %[temp1], 4(%[v1]) \n\t" + "lwc1 %[temp2], 8(%[v1]) \n\t" + "lwc1 %[temp3], 12(%[v1]) \n\t" + "lwc1 %[temp4], 0(%[v2]) \n\t" + "lwc1 %[temp5], 4(%[v2]) \n\t" + "lwc1 %[temp6], 8(%[v2]) \n\t" + "lwc1 %[temp7], 12(%[v2]) \n\t" + "beq %[pom], $zero, 2f \n\t" + "1: \n\t" + "sub.s %[temp8], %[temp0], %[temp4] \n\t" + "add.s %[temp9], %[temp0], %[temp4] \n\t" + "sub.s %[temp10], %[temp1], %[temp5] \n\t" + "add.s %[temp11], %[temp1], %[temp5] \n\t" + "sub.s %[temp12], %[temp2], %[temp6] \n\t" + "add.s %[temp13], %[temp2], %[temp6] \n\t" + "sub.s %[temp14], %[temp3], %[temp7] \n\t" + "add.s %[temp15], %[temp3], %[temp7] \n\t" + PTR_ADDIU "%[v1], %[v1], 16 \n\t" + PTR_ADDIU "%[v2], %[v2], 16 \n\t" + "addiu %[pom], %[pom], -1 \n\t" + "lwc1 %[temp0], 0(%[v1]) \n\t" + "lwc1 %[temp1], 4(%[v1]) \n\t" + "lwc1 %[temp2], 8(%[v1]) \n\t" + "lwc1 %[temp3], 12(%[v1]) \n\t" + "lwc1 %[temp4], 0(%[v2]) \n\t" + "lwc1 %[temp5], 4(%[v2]) \n\t" + "lwc1 %[temp6], 8(%[v2]) \n\t" + "lwc1 %[temp7], 12(%[v2]) \n\t" + "swc1 %[temp9], -16(%[v1]) \n\t" + "swc1 %[temp8], -16(%[v2]) \n\t" + "swc1 %[temp11], -12(%[v1]) \n\t" + "swc1 %[temp10], -12(%[v2]) \n\t" + "swc1 %[temp13], -8(%[v1]) \n\t" + "swc1 %[temp12], -8(%[v2]) \n\t" + "swc1 %[temp15], -4(%[v1]) \n\t" + "swc1 %[temp14], -4(%[v2]) \n\t" + "bgtz %[pom], 1b \n\t" + "2: \n\t" + "sub.s %[temp8], %[temp0], %[temp4] \n\t" + "add.s %[temp9], %[temp0], %[temp4] \n\t" + "sub.s %[temp10], %[temp1], %[temp5] \n\t" + "add.s %[temp11], %[temp1], %[temp5] \n\t" + "sub.s %[temp12], %[temp2], %[temp6] \n\t" + "add.s %[temp13], %[temp2], %[temp6] \n\t" + "sub.s %[temp14], %[temp3], %[temp7] \n\t" + "add.s %[temp15], %[temp3], %[temp7] \n\t" + "swc1 %[temp9], 0(%[v1]) \n\t" + "swc1 %[temp8], 0(%[v2]) \n\t" + "swc1 %[temp11], 4(%[v1]) \n\t" + "swc1 %[temp10], 4(%[v2]) \n\t" + "swc1 %[temp13], 8(%[v1]) \n\t" + "swc1 %[temp12], 8(%[v2]) \n\t" + "swc1 %[temp15], 12(%[v1]) \n\t" + "swc1 %[temp14], 12(%[v2]) \n\t" + + : [v1]"+r"(v1), [v2]"+r"(v2), [pom]"+r"(pom), [temp0] "=&f" (temp0), + [temp1]"=&f"(temp1), [temp2]"=&f"(temp2), [temp3]"=&f"(temp3), + [temp4]"=&f"(temp4), [temp5]"=&f"(temp5), [temp6]"=&f"(temp6), + [temp7]"=&f"(temp7), [temp8]"=&f"(temp8), [temp9]"=&f"(temp9), + [temp10]"=&f"(temp10), [temp11]"=&f"(temp11), [temp12]"=&f"(temp12), + [temp13]"=&f"(temp13), [temp14]"=&f"(temp14), [temp15]"=&f"(temp15) + : + : "memory" + ); +} + +static void vector_fmul_reverse_mips(float *dst, const float *src0, const float *src1, int len){ + int i; + float temp0, temp1, temp2, temp3, temp4, temp5, temp6, temp7; + src1 += len-1; + + for(i=0; i<(len>>2); i++) + { + /* loop unrolled 4 times */ + __asm__ volatile( + "lwc1 %[temp0], 0(%[src0]) \n\t" + "lwc1 %[temp1], 0(%[src1]) \n\t" + "lwc1 %[temp2], 4(%[src0]) \n\t" + "lwc1 %[temp3], -4(%[src1]) \n\t" + "lwc1 %[temp4], 8(%[src0]) \n\t" + "lwc1 %[temp5], -8(%[src1]) \n\t" + "lwc1 %[temp6], 12(%[src0]) \n\t" + "lwc1 %[temp7], -12(%[src1]) \n\t" + "mul.s %[temp0], %[temp1], %[temp0] \n\t" + "mul.s %[temp2], %[temp3], %[temp2] \n\t" + "mul.s %[temp4], %[temp5], %[temp4] \n\t" + "mul.s %[temp6], %[temp7], %[temp6] \n\t" + PTR_ADDIU "%[src0], %[src0], 16 \n\t" + PTR_ADDIU "%[src1], %[src1], -16 \n\t" + PTR_ADDIU "%[dst], %[dst], 16 \n\t" + "swc1 %[temp0], -16(%[dst]) \n\t" + "swc1 %[temp2], -12(%[dst]) \n\t" + "swc1 %[temp4], -8(%[dst]) \n\t" + "swc1 %[temp6], -4(%[dst]) \n\t" + + : [dst]"+r"(dst), [src0]"+r"(src0), [src1]"+r"(src1), + [temp0]"=&f"(temp0), [temp1]"=&f"(temp1),[temp2]"=&f"(temp2), + [temp3]"=&f"(temp3), [temp4]"=&f"(temp4), [temp5]"=&f"(temp5), + [temp6]"=&f"(temp6), [temp7]"=&f"(temp7) + : + : "memory" + ); + } +} +#endif /* !HAVE_MIPS32R6 && !HAVE_MIPS64R6 */ +#endif /* HAVE_INLINE_ASM && HAVE_MIPSFPU */ + +void ff_float_dsp_init_mips(AVFloatDSPContext *fdsp) { +#if HAVE_INLINE_ASM && HAVE_MIPSFPU +#if !HAVE_MIPS32R6 && !HAVE_MIPS64R6 + fdsp->vector_fmul = vector_fmul_mips; + fdsp->vector_fmul_scalar = vector_fmul_scalar_mips; + fdsp->vector_fmul_window = vector_fmul_window_mips; + fdsp->butterflies_float = butterflies_float_mips; + fdsp->vector_fmul_reverse = vector_fmul_reverse_mips; +#endif /* !HAVE_MIPS32R6 && !HAVE_MIPS64R6 */ +#endif /* HAVE_INLINE_ASM && HAVE_MIPSFPU */ +} diff --git a/libavutil/mips/generic_macros_msa.h b/libavutil/mips/generic_macros_msa.h new file mode 100644 index 0000000000..b1d18dd9d3 --- /dev/null +++ b/libavutil/mips/generic_macros_msa.h @@ -0,0 +1,2959 @@ +/* + * Copyright (c) 2015 Manojkumar Bhosale (Manojkumar.Bhosale@imgtec.com) + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#ifndef AVUTIL_MIPS_GENERIC_MACROS_MSA_H +#define AVUTIL_MIPS_GENERIC_MACROS_MSA_H + +#include <stdint.h> +#include <msa.h> + +#define ALIGNMENT 16 +#define ALLOC_ALIGNED(align) __attribute__ ((aligned((align) << 1))) + +#define LD_B(RTYPE, psrc) *((RTYPE *)(psrc)) +#define LD_UB(...) LD_B(v16u8, __VA_ARGS__) +#define LD_SB(...) LD_B(v16i8, __VA_ARGS__) + +#define LD_H(RTYPE, psrc) *((RTYPE *)(psrc)) +#define LD_UH(...) LD_H(v8u16, __VA_ARGS__) +#define LD_SH(...) LD_H(v8i16, __VA_ARGS__) + +#define LD_W(RTYPE, psrc) *((RTYPE *)(psrc)) +#define LD_UW(...) LD_W(v4u32, __VA_ARGS__) +#define LD_SW(...) LD_W(v4i32, __VA_ARGS__) + +#define ST_B(RTYPE, in, pdst) *((RTYPE *)(pdst)) = (in) +#define ST_UB(...) ST_B(v16u8, __VA_ARGS__) +#define ST_SB(...) ST_B(v16i8, __VA_ARGS__) + +#define ST_H(RTYPE, in, pdst) *((RTYPE *)(pdst)) = (in) +#define ST_UH(...) ST_H(v8u16, __VA_ARGS__) +#define ST_SH(...) ST_H(v8i16, __VA_ARGS__) + +#define ST_W(RTYPE, in, pdst) *((RTYPE *)(pdst)) = (in) +#define ST_UW(...) ST_W(v4u32, __VA_ARGS__) +#define ST_SW(...) ST_W(v4i32, __VA_ARGS__) + +#if (__mips_isa_rev >= 6) + #define LW(psrc) \ + ( { \ + uint8_t *psrc_m = (uint8_t *) (psrc); \ + uint32_t val_m; \ + \ + __asm__ volatile ( \ + "lw %[val_m], %[psrc_m] \n\t" \ + \ + : [val_m] "=r" (val_m) \ + : [psrc_m] "m" (*psrc_m) \ + ); \ + \ + val_m; \ + } ) + + #if (__mips == 64) + #define LD(psrc) \ + ( { \ + uint8_t *psrc_m = (uint8_t *) (psrc); \ + uint64_t val_m = 0; \ + \ + __asm__ volatile ( \ + "ld %[val_m], %[psrc_m] \n\t" \ + \ + : [val_m] "=r" (val_m) \ + : [psrc_m] "m" (*psrc_m) \ + ); \ + \ + val_m; \ + } ) + #else // !(__mips == 64) + #define LD(psrc) \ + ( { \ + uint8_t *psrc_m = (uint8_t *) (psrc); \ + uint32_t val0_m, val1_m; \ + uint64_t val_m = 0; \ + \ + val0_m = LW(psrc_m); \ + val1_m = LW(psrc_m + 4); \ + \ + val_m = (uint64_t) (val1_m); \ + val_m = (uint64_t) ((val_m << 32) & 0xFFFFFFFF00000000); \ + val_m = (uint64_t) (val_m | (uint64_t) val0_m); \ + \ + val_m; \ + } ) + #endif // (__mips == 64) + + #define SH(val, pdst) \ + { \ + uint8_t *pdst_m = (uint8_t *) (pdst); \ + uint16_t val_m = (val); \ + \ + __asm__ volatile ( \ + "sh %[val_m], %[pdst_m] \n\t" \ + \ + : [pdst_m] "=m" (*pdst_m) \ + : [val_m] "r" (val_m) \ + ); \ + } + + #define SW(val, pdst) \ + { \ + uint8_t *pdst_m = (uint8_t *) (pdst); \ + uint32_t val_m = (val); \ + \ + __asm__ volatile ( \ + "sw %[val_m], %[pdst_m] \n\t" \ + \ + : [pdst_m] "=m" (*pdst_m) \ + : [val_m] "r" (val_m) \ + ); \ + } + + #define SD(val, pdst) \ + { \ + uint8_t *pdst_m = (uint8_t *) (pdst); \ + uint64_t val_m = (val); \ + \ + __asm__ volatile ( \ + "sd %[val_m], %[pdst_m] \n\t" \ + \ + : [pdst_m] "=m" (*pdst_m) \ + : [val_m] "r" (val_m) \ + ); \ + } +#else // !(__mips_isa_rev >= 6) + #define LW(psrc) \ + ( { \ + uint8_t *psrc_m = (uint8_t *) (psrc); \ + uint32_t val_m; \ + \ + __asm__ volatile ( \ + "ulw %[val_m], %[psrc_m] \n\t" \ + \ + : [val_m] "=r" (val_m) \ + : [psrc_m] "m" (*psrc_m) \ + ); \ + \ + val_m; \ + } ) + + #if (__mips == 64) + #define LD(psrc) \ + ( { \ + uint8_t *psrc_m = (uint8_t *) (psrc); \ + uint64_t val_m = 0; \ + \ + __asm__ volatile ( \ + "uld %[val_m], %[psrc_m] \n\t" \ + \ + : [val_m] "=r" (val_m) \ + : [psrc_m] "m" (*psrc_m) \ + ); \ + \ + val_m; \ + } ) + #else // !(__mips == 64) + #define LD(psrc) \ + ( { \ + uint8_t *psrc_m1 = (uint8_t *) (psrc); \ + uint32_t val0_m, val1_m; \ + uint64_t val_m = 0; \ + \ + val0_m = LW(psrc_m1); \ + val1_m = LW(psrc_m1 + 4); \ + \ + val_m = (uint64_t) (val1_m); \ + val_m = (uint64_t) ((val_m << 32) & 0xFFFFFFFF00000000); \ + val_m = (uint64_t) (val_m | (uint64_t) val0_m); \ + \ + val_m; \ + } ) + #endif // (__mips == 64) + + #define SH(val, pdst) \ + { \ + uint8_t *pdst_m = (uint8_t *) (pdst); \ + uint16_t val_m = (val); \ + \ + __asm__ volatile ( \ + "ush %[val_m], %[pdst_m] \n\t" \ + \ + : [pdst_m] "=m" (*pdst_m) \ + : [val_m] "r" (val_m) \ + ); \ + } + + #define SW(val, pdst) \ + { \ + uint8_t *pdst_m = (uint8_t *) (pdst); \ + uint32_t val_m = (val); \ + \ + __asm__ volatile ( \ + "usw %[val_m], %[pdst_m] \n\t" \ + \ + : [pdst_m] "=m" (*pdst_m) \ + : [val_m] "r" (val_m) \ + ); \ + } + + #define SD(val, pdst) \ + { \ + uint8_t *pdst_m1 = (uint8_t *) (pdst); \ + uint32_t val0_m, val1_m; \ + \ + val0_m = (uint32_t) ((val) & 0x00000000FFFFFFFF); \ + val1_m = (uint32_t) (((val) >> 32) & 0x00000000FFFFFFFF); \ + \ + SW(val0_m, pdst_m1); \ + SW(val1_m, pdst_m1 + 4); \ + } +#endif // (__mips_isa_rev >= 6) + +/* Description : Load 4 words with stride + Arguments : Inputs - psrc (source pointer to load from) + - stride + Outputs - out0, out1, out2, out3 + Details : Loads word in 'out0' from (psrc) + Loads word in 'out1' from (psrc + stride) + Loads word in 'out2' from (psrc + 2 * stride) + Loads word in 'out3' from (psrc + 3 * stride) +*/ +#define LW4(psrc, stride, out0, out1, out2, out3) \ +{ \ + out0 = LW((psrc)); \ + out1 = LW((psrc) + stride); \ + out2 = LW((psrc) + 2 * stride); \ + out3 = LW((psrc) + 3 * stride); \ +} + +/* Description : Load double words with stride + Arguments : Inputs - psrc (source pointer to load from) + - stride + Outputs - out0, out1 + Details : Loads double word in 'out0' from (psrc) + Loads double word in 'out1' from (psrc + stride) +*/ +#define LD2(psrc, stride, out0, out1) \ +{ \ + out0 = LD((psrc)); \ + out1 = LD((psrc) + stride); \ +} +#define LD4(psrc, stride, out0, out1, out2, out3) \ +{ \ + LD2((psrc), stride, out0, out1); \ + LD2((psrc) + 2 * stride, stride, out2, out3); \ +} + +/* Description : Store 4 words with stride + Arguments : Inputs - in0, in1, in2, in3, pdst, stride + Details : Stores word from 'in0' to (pdst) + Stores word from 'in1' to (pdst + stride) + Stores word from 'in2' to (pdst + 2 * stride) + Stores word from 'in3' to (pdst + 3 * stride) +*/ +#define SW4(in0, in1, in2, in3, pdst, stride) \ +{ \ + SW(in0, (pdst)) \ + SW(in1, (pdst) + stride); \ + SW(in2, (pdst) + 2 * stride); \ + SW(in3, (pdst) + 3 * stride); \ +} + +/* Description : Store 4 double words with stride + Arguments : Inputs - in0, in1, in2, in3, pdst, stride + Details : Stores double word from 'in0' to (pdst) + Stores double word from 'in1' to (pdst + stride) + Stores double word from 'in2' to (pdst + 2 * stride) + Stores double word from 'in3' to (pdst + 3 * stride) +*/ +#define SD4(in0, in1, in2, in3, pdst, stride) \ +{ \ + SD(in0, (pdst)) \ + SD(in1, (pdst) + stride); \ + SD(in2, (pdst) + 2 * stride); \ + SD(in3, (pdst) + 3 * stride); \ +} + +/* Description : Load vectors with 16 byte elements with stride + Arguments : Inputs - psrc (source pointer to load from) + - stride + Outputs - out0, out1 + Return Type - as per RTYPE + Details : Loads 16 byte elements in 'out0' from (psrc) + Loads 16 byte elements in 'out1' from (psrc + stride) +*/ +#define LD_B2(RTYPE, psrc, stride, out0, out1) \ +{ \ + out0 = LD_B(RTYPE, (psrc)); \ + out1 = LD_B(RTYPE, (psrc) + stride); \ +} +#define LD_UB2(...) LD_B2(v16u8, __VA_ARGS__) +#define LD_SB2(...) LD_B2(v16i8, __VA_ARGS__) + +#define LD_B3(RTYPE, psrc, stride, out0, out1, out2) \ +{ \ + LD_B2(RTYPE, (psrc), stride, out0, out1); \ + out2 = LD_B(RTYPE, (psrc) + 2 * stride); \ +} +#define LD_UB3(...) LD_B3(v16u8, __VA_ARGS__) +#define LD_SB3(...) LD_B3(v16i8, __VA_ARGS__) + +#define LD_B4(RTYPE, psrc, stride, out0, out1, out2, out3) \ +{ \ + LD_B2(RTYPE, (psrc), stride, out0, out1); \ + LD_B2(RTYPE, (psrc) + 2 * stride , stride, out2, out3); \ +} +#define LD_UB4(...) LD_B4(v16u8, __VA_ARGS__) +#define LD_SB4(...) LD_B4(v16i8, __VA_ARGS__) + +#define LD_B5(RTYPE, psrc, stride, out0, out1, out2, out3, out4) \ +{ \ + LD_B4(RTYPE, (psrc), stride, out0, out1, out2, out3); \ + out4 = LD_B(RTYPE, (psrc) + 4 * stride); \ +} +#define LD_UB5(...) LD_B5(v16u8, __VA_ARGS__) +#define LD_SB5(...) LD_B5(v16i8, __VA_ARGS__) + +#define LD_B6(RTYPE, psrc, stride, out0, out1, out2, out3, out4, out5) \ +{ \ + LD_B4(RTYPE, (psrc), stride, out0, out1, out2, out3); \ + LD_B2(RTYPE, (psrc) + 4 * stride, stride, out4, out5); \ +} +#define LD_UB6(...) LD_B6(v16u8, __VA_ARGS__) +#define LD_SB6(...) LD_B6(v16i8, __VA_ARGS__) + +#define LD_B7(RTYPE, psrc, stride, \ + out0, out1, out2, out3, out4, out5, out6) \ +{ \ + LD_B5(RTYPE, (psrc), stride, out0, out1, out2, out3, out4); \ + LD_B2(RTYPE, (psrc) + 5 * stride, stride, out5, out6); \ +} +#define LD_UB7(...) LD_B7(v16u8, __VA_ARGS__) +#define LD_SB7(...) LD_B7(v16i8, __VA_ARGS__) + +#define LD_B8(RTYPE, psrc, stride, \ + out0, out1, out2, out3, out4, out5, out6, out7) \ +{ \ + LD_B4(RTYPE, (psrc), stride, out0, out1, out2, out3); \ + LD_B4(RTYPE, (psrc) + 4 * stride, stride, out4, out5, out6, out7); \ +} +#define LD_UB8(...) LD_B8(v16u8, __VA_ARGS__) +#define LD_SB8(...) LD_B8(v16i8, __VA_ARGS__) + +/* Description : Load vectors with 8 halfword elements with stride + Arguments : Inputs - psrc (source pointer to load from) + - stride + Outputs - out0, out1 + Details : Loads 8 halfword elements in 'out0' from (psrc) + Loads 8 halfword elements in 'out1' from (psrc + stride) +*/ +#define LD_H2(RTYPE, psrc, stride, out0, out1) \ +{ \ + out0 = LD_H(RTYPE, (psrc)); \ + out1 = LD_H(RTYPE, (psrc) + (stride)); \ +} +#define LD_UH2(...) LD_H2(v8u16, __VA_ARGS__) +#define LD_SH2(...) LD_H2(v8i16, __VA_ARGS__) + +#define LD_H4(RTYPE, psrc, stride, out0, out1, out2, out3) \ +{ \ + LD_H2(RTYPE, (psrc), stride, out0, out1); \ + LD_H2(RTYPE, (psrc) + 2 * stride, stride, out2, out3); \ +} +#define LD_UH4(...) LD_H4(v8u16, __VA_ARGS__) +#define LD_SH4(...) LD_H4(v8i16, __VA_ARGS__) + +#define LD_H6(RTYPE, psrc, stride, out0, out1, out2, out3, out4, out5) \ +{ \ + LD_H4(RTYPE, (psrc), stride, out0, out1, out2, out3); \ + LD_H2(RTYPE, (psrc) + 4 * stride, stride, out4, out5); \ +} +#define LD_UH6(...) LD_H6(v8u16, __VA_ARGS__) +#define LD_SH6(...) LD_H6(v8i16, __VA_ARGS__) + +#define LD_H8(RTYPE, psrc, stride, \ + out0, out1, out2, out3, out4, out5, out6, out7) \ +{ \ + LD_H4(RTYPE, (psrc), stride, out0, out1, out2, out3); \ + LD_H4(RTYPE, (psrc) + 4 * stride, stride, out4, out5, out6, out7); \ +} +#define LD_UH8(...) LD_H8(v8u16, __VA_ARGS__) +#define LD_SH8(...) LD_H8(v8i16, __VA_ARGS__) + +#define LD_H16(RTYPE, psrc, stride, \ + out0, out1, out2, out3, out4, out5, out6, out7, \ + out8, out9, out10, out11, out12, out13, out14, out15) \ +{ \ + LD_H8(RTYPE, (psrc), stride, \ + out0, out1, out2, out3, out4, out5, out6, out7); \ + LD_H8(RTYPE, (psrc) + 8 * stride, stride, \ + out8, out9, out10, out11, out12, out13, out14, out15); \ +} +#define LD_SH16(...) LD_H16(v8i16, __VA_ARGS__) + +/* Description : Load as 4x4 block of signed halfword elements from 1D source + data into 4 vectors (Each vector with 4 signed halfwords) + Arguments : Inputs - psrc + Outputs - out0, out1, out2, out3 +*/ +#define LD4x4_SH(psrc, out0, out1, out2, out3) \ +{ \ + out0 = LD_SH(psrc); \ + out2 = LD_SH(psrc + 8); \ + out1 = (v8i16) __msa_ilvl_d((v2i64) out0, (v2i64) out0); \ + out3 = (v8i16) __msa_ilvl_d((v2i64) out2, (v2i64) out2); \ +} + +/* Description : Load 2 vectors of signed word elements with stride + Arguments : Inputs - psrc (source pointer to load from) + - stride + Outputs - out0, out1 + Return Type - signed word +*/ +#define LD_SW2(psrc, stride, out0, out1) \ +{ \ + out0 = LD_SW((psrc)); \ + out1 = LD_SW((psrc) + stride); \ +} + +/* Description : Store vectors of 16 byte elements with stride + Arguments : Inputs - in0, in1, stride + Outputs - pdst (destination pointer to store to) + Details : Stores 16 byte elements from 'in0' to (pdst) + Stores 16 byte elements from 'in1' to (pdst + stride) +*/ +#define ST_B2(RTYPE, in0, in1, pdst, stride) \ +{ \ + ST_B(RTYPE, in0, (pdst)); \ + ST_B(RTYPE, in1, (pdst) + stride); \ +} +#define ST_UB2(...) ST_B2(v16u8, __VA_ARGS__) +#define ST_SB2(...) ST_B2(v16i8, __VA_ARGS__) + +#define ST_B4(RTYPE, in0, in1, in2, in3, pdst, stride) \ +{ \ + ST_B2(RTYPE, in0, in1, (pdst), stride); \ + ST_B2(RTYPE, in2, in3, (pdst) + 2 * stride, stride); \ +} +#define ST_UB4(...) ST_B4(v16u8, __VA_ARGS__) +#define ST_SB4(...) ST_B4(v16i8, __VA_ARGS__) + +#define ST_B8(RTYPE, in0, in1, in2, in3, in4, in5, in6, in7, \ + pdst, stride) \ +{ \ + ST_B4(RTYPE, in0, in1, in2, in3, pdst, stride); \ + ST_B4(RTYPE, in4, in5, in6, in7, (pdst) + 4 * stride, stride); \ +} +#define ST_UB8(...) ST_B8(v16u8, __VA_ARGS__) + +/* Description : Store vectors of 8 halfword elements with stride + Arguments : Inputs - in0, in1, stride + Outputs - pdst (destination pointer to store to) + Details : Stores 8 halfword elements from 'in0' to (pdst) + Stores 8 halfword elements from 'in1' to (pdst + stride) +*/ +#define ST_H2(RTYPE, in0, in1, pdst, stride) \ +{ \ + ST_H(RTYPE, in0, (pdst)); \ + ST_H(RTYPE, in1, (pdst) + stride); \ +} +#define ST_UH2(...) ST_H2(v8u16, __VA_ARGS__) +#define ST_SH2(...) ST_H2(v8i16, __VA_ARGS__) + +#define ST_H4(RTYPE, in0, in1, in2, in3, pdst, stride) \ +{ \ + ST_H2(RTYPE, in0, in1, (pdst), stride); \ + ST_H2(RTYPE, in2, in3, (pdst) + 2 * stride, stride); \ +} +#define ST_SH4(...) ST_H4(v8i16, __VA_ARGS__) + +#define ST_H6(RTYPE, in0, in1, in2, in3, in4, in5, pdst, stride) \ +{ \ + ST_H4(RTYPE, in0, in1, in2, in3, (pdst), stride); \ + ST_H2(RTYPE, in4, in5, (pdst) + 4 * stride, stride); \ +} +#define ST_SH6(...) ST_H6(v8i16, __VA_ARGS__) + +#define ST_H8(RTYPE, in0, in1, in2, in3, in4, in5, in6, in7, pdst, stride) \ +{ \ + ST_H4(RTYPE, in0, in1, in2, in3, (pdst), stride); \ + ST_H4(RTYPE, in4, in5, in6, in7, (pdst) + 4 * stride, stride); \ +} +#define ST_SH8(...) ST_H8(v8i16, __VA_ARGS__) + +/* Description : Store vectors of word elements with stride + Arguments : Inputs - in0, in1, stride + Outputs - pdst (destination pointer to store to) + Return Type - signed word + Details : Stores 4 word elements from 'in0' to (pdst) + Stores 4 word elements from 'in1' to (pdst + stride) +*/ +#define ST_SW2(in0, in1, pdst, stride) \ +{ \ + ST_SW(in0, (pdst)); \ + ST_SW(in1, (pdst) + stride); \ +} +#define ST_SW8(in0, in1, in2, in3, in4, in5, in6, in7, \ + pdst, stride) \ +{ \ + ST_SW2(in0, in1, (pdst), stride); \ + ST_SW2(in2, in3, (pdst) + 2 * stride, stride); \ + ST_SW2(in4, in5, (pdst) + 4 * stride, stride); \ + ST_SW2(in6, in7, (pdst) + 6 * stride, stride); \ +} + +/* Description : Store as 2x4 byte block to destination memory from input vector + Arguments : Inputs - in, stidx, pdst, stride + Return Type - unsigned byte + Details : Index stidx halfword element from 'in' vector is copied and + stored on first line + Index stidx+1 halfword element from 'in' vector is copied and + stored on second line + Index stidx+2 halfword element from 'in' vector is copied and + stored on third line + Index stidx+3 halfword element from 'in' vector is copied and + stored on fourth line +*/ +#define ST2x4_UB(in, stidx, pdst, stride) \ +{ \ + uint16_t out0_m, out1_m, out2_m, out3_m; \ + uint8_t *pblk_2x4_m = (uint8_t *) (pdst); \ + \ + out0_m = __msa_copy_u_h((v8i16) in, (stidx)); \ + out1_m = __msa_copy_u_h((v8i16) in, (stidx + 1)); \ + out2_m = __msa_copy_u_h((v8i16) in, (stidx + 2)); \ + out3_m = __msa_copy_u_h((v8i16) in, (stidx + 3)); \ + \ + SH(out0_m, pblk_2x4_m); \ + SH(out1_m, pblk_2x4_m + stride); \ + SH(out2_m, pblk_2x4_m + 2 * stride); \ + SH(out3_m, pblk_2x4_m + 3 * stride); \ +} + +/* Description : Store as 4x2 byte block to destination memory from input vector + Arguments : Inputs - in, pdst, stride + Return Type - unsigned byte + Details : Index 0 word element from input vector is copied and stored + on first line + Index 1 word element from input vector is copied and stored + on second line +*/ +#define ST4x2_UB(in, pdst, stride) \ +{ \ + uint32_t out0_m, out1_m; \ + uint8_t *pblk_4x2_m = (uint8_t *) (pdst); \ + \ + out0_m = __msa_copy_u_w((v4i32) in, 0); \ + out1_m = __msa_copy_u_w((v4i32) in, 1); \ + \ + SW(out0_m, pblk_4x2_m); \ + SW(out1_m, pblk_4x2_m + stride); \ +} + +/* Description : Store as 4x4 byte block to destination memory from input vector + Arguments : Inputs - in0, in1, pdst, stride + Return Type - unsigned byte + Details : Idx0 word element from input vector 'in0' is copied and stored + on first line + Idx1 word element from input vector 'in0' is copied and stored + on second line + Idx2 word element from input vector 'in1' is copied and stored + on third line + Idx3 word element from input vector 'in1' is copied and stored + on fourth line +*/ +#define ST4x4_UB(in0, in1, idx0, idx1, idx2, idx3, pdst, stride) \ +{ \ + uint32_t out0_m, out1_m, out2_m, out3_m; \ + uint8_t *pblk_4x4_m = (uint8_t *) (pdst); \ + \ + out0_m = __msa_copy_u_w((v4i32) in0, idx0); \ + out1_m = __msa_copy_u_w((v4i32) in0, idx1); \ + out2_m = __msa_copy_u_w((v4i32) in1, idx2); \ + out3_m = __msa_copy_u_w((v4i32) in1, idx3); \ + \ + SW4(out0_m, out1_m, out2_m, out3_m, pblk_4x4_m, stride); \ +} +#define ST4x8_UB(in0, in1, pdst, stride) \ +{ \ + uint8_t *pblk_4x8 = (uint8_t *) (pdst); \ + \ + ST4x4_UB(in0, in0, 0, 1, 2, 3, pblk_4x8, stride); \ + ST4x4_UB(in1, in1, 0, 1, 2, 3, pblk_4x8 + 4 * stride, stride); \ +} + +/* Description : Store as 6x4 byte block to destination memory from input + vectors + Arguments : Inputs - in0, in1, pdst, stride + Return Type - unsigned byte + Details : Index 0 word element from input vector 'in0' is copied and + stored on first line followed by index 2 halfword element + Index 2 word element from input vector 'in0' is copied and + stored on second line followed by index 2 halfword element + Index 0 word element from input vector 'in1' is copied and + stored on third line followed by index 2 halfword element + Index 2 word element from input vector 'in1' is copied and + stored on fourth line followed by index 2 halfword element +*/ +#define ST6x4_UB(in0, in1, pdst, stride) \ +{ \ + uint32_t out0_m, out1_m, out2_m, out3_m; \ + uint16_t out4_m, out5_m, out6_m, out7_m; \ + uint8_t *pblk_6x4_m = (uint8_t *) (pdst); \ + \ + out0_m = __msa_copy_u_w((v4i32) in0, 0); \ + out1_m = __msa_copy_u_w((v4i32) in0, 2); \ + out2_m = __msa_copy_u_w((v4i32) in1, 0); \ + out3_m = __msa_copy_u_w((v4i32) in1, 2); \ + \ + out4_m = __msa_copy_u_h((v8i16) in0, 2); \ + out5_m = __msa_copy_u_h((v8i16) in0, 6); \ + out6_m = __msa_copy_u_h((v8i16) in1, 2); \ + out7_m = __msa_copy_u_h((v8i16) in1, 6); \ + \ + SW(out0_m, pblk_6x4_m); \ + SH(out4_m, (pblk_6x4_m + 4)); \ + pblk_6x4_m += stride; \ + SW(out1_m, pblk_6x4_m); \ + SH(out5_m, (pblk_6x4_m + 4)); \ + pblk_6x4_m += stride; \ + SW(out2_m, pblk_6x4_m); \ + SH(out6_m, (pblk_6x4_m + 4)); \ + pblk_6x4_m += stride; \ + SW(out3_m, pblk_6x4_m); \ + SH(out7_m, (pblk_6x4_m + 4)); \ +} + +/* Description : Store as 8x1 byte block to destination memory from input vector + Arguments : Inputs - in, pdst + Details : Index 0 double word element from input vector 'in' is copied + and stored to destination memory at (pdst) +*/ +#define ST8x1_UB(in, pdst) \ +{ \ + uint64_t out0_m; \ + out0_m = __msa_copy_u_d((v2i64) in, 0); \ + SD(out0_m, pdst); \ +} + +/* Description : Store as 8x2 byte block to destination memory from input vector + Arguments : Inputs - in, pdst, stride + Details : Index 0 double word element from input vector 'in' is copied + and stored to destination memory at (pdst) + Index 1 double word element from input vector 'in' is copied + and stored to destination memory at (pdst + stride) +*/ +#define ST8x2_UB(in, pdst, stride) \ +{ \ + uint64_t out0_m, out1_m; \ + uint8_t *pblk_8x2_m = (uint8_t *) (pdst); \ + \ + out0_m = __msa_copy_u_d((v2i64) in, 0); \ + out1_m = __msa_copy_u_d((v2i64) in, 1); \ + \ + SD(out0_m, pblk_8x2_m); \ + SD(out1_m, pblk_8x2_m + stride); \ +} + +/* Description : Store as 8x4 byte block to destination memory from input + vectors + Arguments : Inputs - in0, in1, pdst, stride + Details : Index 0 double word element from input vector 'in0' is copied + and stored to destination memory at (pblk_8x4_m) + Index 1 double word element from input vector 'in0' is copied + and stored to destination memory at (pblk_8x4_m + stride) + Index 0 double word element from input vector 'in1' is copied + and stored to destination memory at (pblk_8x4_m + 2 * stride) + Index 1 double word element from input vector 'in1' is copied + and stored to destination memory at (pblk_8x4_m + 3 * stride) +*/ +#define ST8x4_UB(in0, in1, pdst, stride) \ +{ \ + uint64_t out0_m, out1_m, out2_m, out3_m; \ + uint8_t *pblk_8x4_m = (uint8_t *) (pdst); \ + \ + out0_m = __msa_copy_u_d((v2i64) in0, 0); \ + out1_m = __msa_copy_u_d((v2i64) in0, 1); \ + out2_m = __msa_copy_u_d((v2i64) in1, 0); \ + out3_m = __msa_copy_u_d((v2i64) in1, 1); \ + \ + SD4(out0_m, out1_m, out2_m, out3_m, pblk_8x4_m, stride); \ +} +#define ST8x8_UB(in0, in1, in2, in3, pdst, stride) \ +{ \ + uint8_t *pblk_8x8_m = (uint8_t *) (pdst); \ + \ + ST8x4_UB(in0, in1, pblk_8x8_m, stride); \ + ST8x4_UB(in2, in3, pblk_8x8_m + 4 * stride, stride); \ +} +#define ST12x4_UB(in0, in1, in2, pdst, stride) \ +{ \ + uint8_t *pblk_12x4_m = (uint8_t *) (pdst); \ + \ + /* left 8x4 */ \ + ST8x4_UB(in0, in1, pblk_12x4_m, stride); \ + /* right 4x4 */ \ + ST4x4_UB(in2, in2, 0, 1, 2, 3, pblk_12x4_m + 8, stride); \ +} + +/* Description : Store as 12x8 byte block to destination memory from + input vectors + Arguments : Inputs - in0, in1, in2, in3, in4, in5, in6, in7, pdst, stride + Details : Index 0 double word element from input vector 'in0' is copied + and stored to destination memory at (pblk_12x8_m) followed by + index 2 word element from same input vector 'in0' at + (pblk_12x8_m + 8) + Similar to remaining lines +*/ +#define ST12x8_UB(in0, in1, in2, in3, in4, in5, in6, in7, pdst, stride) \ +{ \ + uint64_t out0_m, out1_m, out2_m, out3_m; \ + uint64_t out4_m, out5_m, out6_m, out7_m; \ + uint32_t out8_m, out9_m, out10_m, out11_m; \ + uint32_t out12_m, out13_m, out14_m, out15_m; \ + uint8_t *pblk_12x8_m = (uint8_t *) (pdst); \ + \ + out0_m = __msa_copy_u_d((v2i64) in0, 0); \ + out1_m = __msa_copy_u_d((v2i64) in1, 0); \ + out2_m = __msa_copy_u_d((v2i64) in2, 0); \ + out3_m = __msa_copy_u_d((v2i64) in3, 0); \ + out4_m = __msa_copy_u_d((v2i64) in4, 0); \ + out5_m = __msa_copy_u_d((v2i64) in5, 0); \ + out6_m = __msa_copy_u_d((v2i64) in6, 0); \ + out7_m = __msa_copy_u_d((v2i64) in7, 0); \ + \ + out8_m = __msa_copy_u_w((v4i32) in0, 2); \ + out9_m = __msa_copy_u_w((v4i32) in1, 2); \ + out10_m = __msa_copy_u_w((v4i32) in2, 2); \ + out11_m = __msa_copy_u_w((v4i32) in3, 2); \ + out12_m = __msa_copy_u_w((v4i32) in4, 2); \ + out13_m = __msa_copy_u_w((v4i32) in5, 2); \ + out14_m = __msa_copy_u_w((v4i32) in6, 2); \ + out15_m = __msa_copy_u_w((v4i32) in7, 2); \ + \ + SD(out0_m, pblk_12x8_m); \ + SW(out8_m, pblk_12x8_m + 8); \ + pblk_12x8_m += stride; \ + SD(out1_m, pblk_12x8_m); \ + SW(out9_m, pblk_12x8_m + 8); \ + pblk_12x8_m += stride; \ + SD(out2_m, pblk_12x8_m); \ + SW(out10_m, pblk_12x8_m + 8); \ + pblk_12x8_m += stride; \ + SD(out3_m, pblk_12x8_m); \ + SW(out11_m, pblk_12x8_m + 8); \ + pblk_12x8_m += stride; \ + SD(out4_m, pblk_12x8_m); \ + SW(out12_m, pblk_12x8_m + 8); \ + pblk_12x8_m += stride; \ + SD(out5_m, pblk_12x8_m); \ + SW(out13_m, pblk_12x8_m + 8); \ + pblk_12x8_m += stride; \ + SD(out6_m, pblk_12x8_m); \ + SW(out14_m, pblk_12x8_m + 8); \ + pblk_12x8_m += stride; \ + SD(out7_m, pblk_12x8_m); \ + SW(out15_m, pblk_12x8_m + 8); \ +} + +/* Description : average with rounding (in0 + in1 + 1) / 2. + Arguments : Inputs - in0, in1, in2, in3, + Outputs - out0, out1 + Return Type - signed byte + Details : Each byte element from 'in0' vector is added with each byte + element from 'in1' vector. The addition of the elements plus 1 + (for rounding) is done unsigned with full precision, + i.e. the result has one extra bit. Unsigned division by 2 + (or logical shift right by one bit) is performed before writing + the result to vector 'out0' + Similar for the pair of 'in2' and 'in3' +*/ +#define AVER_UB2(RTYPE, in0, in1, in2, in3, out0, out1) \ +{ \ + out0 = (RTYPE) __msa_aver_u_b((v16u8) in0, (v16u8) in1); \ + out1 = (RTYPE) __msa_aver_u_b((v16u8) in2, (v16u8) in3); \ +} +#define AVER_UB2_UB(...) AVER_UB2(v16u8, __VA_ARGS__) + +#define AVER_UB4(RTYPE, in0, in1, in2, in3, in4, in5, in6, in7, \ + out0, out1, out2, out3) \ +{ \ + AVER_UB2(RTYPE, in0, in1, in2, in3, out0, out1) \ + AVER_UB2(RTYPE, in4, in5, in6, in7, out2, out3) \ +} +#define AVER_UB4_UB(...) AVER_UB4(v16u8, __VA_ARGS__) + +/* Description : Immediate number of columns to slide with zero + Arguments : Inputs - in0, in1, slide_val + Outputs - out0, out1 + Return Type - as per RTYPE + Details : Byte elements from 'zero_m' vector are slide into 'in0' by + number of elements specified by 'slide_val' +*/ +#define SLDI_B2_0(RTYPE, in0, in1, out0, out1, slide_val) \ +{ \ + v16i8 zero_m = { 0 }; \ + out0 = (RTYPE) __msa_sldi_b((v16i8) zero_m, (v16i8) in0, slide_val); \ + out1 = (RTYPE) __msa_sldi_b((v16i8) zero_m, (v16i8) in1, slide_val); \ +} +#define SLDI_B2_0_UB(...) SLDI_B2_0(v16u8, __VA_ARGS__) +#define SLDI_B2_0_SB(...) SLDI_B2_0(v16i8, __VA_ARGS__) +#define SLDI_B2_0_SW(...) SLDI_B2_0(v4i32, __VA_ARGS__) + +#define SLDI_B3_0(RTYPE, in0, in1, in2, out0, out1, out2, slide_val) \ +{ \ + v16i8 zero_m = { 0 }; \ + SLDI_B2_0(RTYPE, in0, in1, out0, out1, slide_val); \ + out2 = (RTYPE) __msa_sldi_b((v16i8) zero_m, (v16i8) in2, slide_val); \ +} +#define SLDI_B3_0_UB(...) SLDI_B3_0(v16u8, __VA_ARGS__) +#define SLDI_B3_0_SB(...) SLDI_B3_0(v16i8, __VA_ARGS__) + +#define SLDI_B4_0(RTYPE, in0, in1, in2, in3, \ + out0, out1, out2, out3, slide_val) \ +{ \ + SLDI_B2_0(RTYPE, in0, in1, out0, out1, slide_val); \ + SLDI_B2_0(RTYPE, in2, in3, out2, out3, slide_val); \ +} +#define SLDI_B4_0_UB(...) SLDI_B4_0(v16u8, __VA_ARGS__) +#define SLDI_B4_0_SB(...) SLDI_B4_0(v16i8, __VA_ARGS__) +#define SLDI_B4_0_SH(...) SLDI_B4_0(v8i16, __VA_ARGS__) + +/* Description : Immediate number of columns to slide + Arguments : Inputs - in0_0, in0_1, in1_0, in1_1, slide_val + Outputs - out0, out1 + Return Type - as per RTYPE + Details : Byte elements from 'in0_0' vector are slide into 'in1_0' by + number of elements specified by 'slide_val' +*/ +#define SLDI_B2(RTYPE, in0_0, in0_1, in1_0, in1_1, out0, out1, slide_val) \ +{ \ + out0 = (RTYPE) __msa_sldi_b((v16i8) in0_0, (v16i8) in1_0, slide_val); \ + out1 = (RTYPE) __msa_sldi_b((v16i8) in0_1, (v16i8) in1_1, slide_val); \ +} +#define SLDI_B2_UB(...) SLDI_B2(v16u8, __VA_ARGS__) +#define SLDI_B2_SB(...) SLDI_B2(v16i8, __VA_ARGS__) +#define SLDI_B2_SH(...) SLDI_B2(v8i16, __VA_ARGS__) + +#define SLDI_B3(RTYPE, in0_0, in0_1, in0_2, in1_0, in1_1, in1_2, \ + out0, out1, out2, slide_val) \ +{ \ + SLDI_B2(RTYPE, in0_0, in0_1, in1_0, in1_1, out0, out1, slide_val) \ + out2 = (RTYPE) __msa_sldi_b((v16i8) in0_2, (v16i8) in1_2, slide_val); \ +} +#define SLDI_B3_SB(...) SLDI_B3(v16i8, __VA_ARGS__) +#define SLDI_B3_UH(...) SLDI_B3(v8u16, __VA_ARGS__) + +/* Description : Shuffle byte vector elements as per mask vector + Arguments : Inputs - in0, in1, in2, in3, mask0, mask1 + Outputs - out0, out1 + Return Type - as per RTYPE + Details : Selective byte elements from in0 & in1 are copied to out0 as + per control vector mask0 + Selective byte elements from in2 & in3 are copied to out1 as + per control vector mask1 +*/ +#define VSHF_B2(RTYPE, in0, in1, in2, in3, mask0, mask1, out0, out1) \ +{ \ + out0 = (RTYPE) __msa_vshf_b((v16i8) mask0, (v16i8) in1, (v16i8) in0); \ + out1 = (RTYPE) __msa_vshf_b((v16i8) mask1, (v16i8) in3, (v16i8) in2); \ +} +#define VSHF_B2_UB(...) VSHF_B2(v16u8, __VA_ARGS__) +#define VSHF_B2_SB(...) VSHF_B2(v16i8, __VA_ARGS__) +#define VSHF_B2_UH(...) VSHF_B2(v8u16, __VA_ARGS__) +#define VSHF_B2_SH(...) VSHF_B2(v8i16, __VA_ARGS__) + +#define VSHF_B3(RTYPE, in0, in1, in2, in3, in4, in5, mask0, mask1, mask2, \ + out0, out1, out2) \ +{ \ + VSHF_B2(RTYPE, in0, in1, in2, in3, mask0, mask1, out0, out1); \ + out2 = (RTYPE) __msa_vshf_b((v16i8) mask2, (v16i8) in5, (v16i8) in4); \ +} +#define VSHF_B3_SB(...) VSHF_B3(v16i8, __VA_ARGS__) + +#define VSHF_B4(RTYPE, in0, in1, mask0, mask1, mask2, mask3, \ + out0, out1, out2, out3) \ +{ \ + VSHF_B2(RTYPE, in0, in1, in0, in1, mask0, mask1, out0, out1); \ + VSHF_B2(RTYPE, in0, in1, in0, in1, mask2, mask3, out2, out3); \ +} +#define VSHF_B4_SB(...) VSHF_B4(v16i8, __VA_ARGS__) +#define VSHF_B4_SH(...) VSHF_B4(v8i16, __VA_ARGS__) + +/* Description : Shuffle halfword vector elements as per mask vector + Arguments : Inputs - in0, in1, in2, in3, mask0, mask1 + Outputs - out0, out1 + Return Type - as per RTYPE + Details : Selective halfword elements from in0 & in1 are copied to out0 + as per control vector mask0 + Selective halfword elements from in2 & in3 are copied to out1 + as per control vector mask1 +*/ +#define VSHF_H2(RTYPE, in0, in1, in2, in3, mask0, mask1, out0, out1) \ +{ \ + out0 = (RTYPE) __msa_vshf_h((v8i16) mask0, (v8i16) in1, (v8i16) in0); \ + out1 = (RTYPE) __msa_vshf_h((v8i16) mask1, (v8i16) in3, (v8i16) in2); \ +} +#define VSHF_H2_SH(...) VSHF_H2(v8i16, __VA_ARGS__) + +#define VSHF_H3(RTYPE, in0, in1, in2, in3, in4, in5, mask0, mask1, mask2, \ + out0, out1, out2) \ +{ \ + VSHF_H2(RTYPE, in0, in1, in2, in3, mask0, mask1, out0, out1); \ + out2 = (RTYPE) __msa_vshf_h((v8i16) mask2, (v8i16) in5, (v8i16) in4); \ +} +#define VSHF_H3_SH(...) VSHF_H3(v8i16, __VA_ARGS__) + +/* Description : Shuffle byte vector elements as per mask vector + Arguments : Inputs - in0, in1, in2, in3, mask0, mask1 + Outputs - out0, out1 + Return Type - as per RTYPE + Details : Selective byte elements from in0 & in1 are copied to out0 as + per control vector mask0 + Selective byte elements from in2 & in3 are copied to out1 as + per control vector mask1 +*/ +#define VSHF_W2(RTYPE, in0, in1, in2, in3, mask0, mask1, out0, out1) \ +{ \ + out0 = (RTYPE) __msa_vshf_w((v4i32) mask0, (v4i32) in1, (v4i32) in0); \ + out1 = (RTYPE) __msa_vshf_w((v4i32) mask1, (v4i32) in3, (v4i32) in2); \ +} +#define VSHF_W2_SB(...) VSHF_W2(v16i8, __VA_ARGS__) + +/* Description : Dot product of byte vector elements + Arguments : Inputs - mult0, mult1 + cnst0, cnst1 + Outputs - out0, out1 + Return Type - unsigned halfword + Details : Unsigned byte elements from mult0 are multiplied with + unsigned byte elements from cnst0 producing a result + twice the size of input i.e. unsigned halfword. + Then this multiplication results of adjacent odd-even elements + are added together and stored to the out vector + (2 unsigned halfword results) +*/ +#define DOTP_UB2(RTYPE, mult0, mult1, cnst0, cnst1, out0, out1) \ +{ \ + out0 = (RTYPE) __msa_dotp_u_h((v16u8) mult0, (v16u8) cnst0); \ + out1 = (RTYPE) __msa_dotp_u_h((v16u8) mult1, (v16u8) cnst1); \ +} +#define DOTP_UB2_UH(...) DOTP_UB2(v8u16, __VA_ARGS__) + +#define DOTP_UB4(RTYPE, mult0, mult1, mult2, mult3, \ + cnst0, cnst1, cnst2, cnst3, \ + out0, out1, out2, out3) \ +{ \ + DOTP_UB2(RTYPE, mult0, mult1, cnst0, cnst1, out0, out1); \ + DOTP_UB2(RTYPE, mult2, mult3, cnst2, cnst3, out2, out3); \ +} +#define DOTP_UB4_UH(...) DOTP_UB4(v8u16, __VA_ARGS__) + +/* Description : Dot product of byte vector elements + Arguments : Inputs - mult0, mult1 + cnst0, cnst1 + Outputs - out0, out1 + Return Type - signed halfword + Details : Signed byte elements from mult0 are multiplied with + signed byte elements from cnst0 producing a result + twice the size of input i.e. signed halfword. + Then this multiplication results of adjacent odd-even elements + are added together and stored to the out vector + (2 signed halfword results) +*/ +#define DOTP_SB2(RTYPE, mult0, mult1, cnst0, cnst1, out0, out1) \ +{ \ + out0 = (RTYPE) __msa_dotp_s_h((v16i8) mult0, (v16i8) cnst0); \ + out1 = (RTYPE) __msa_dotp_s_h((v16i8) mult1, (v16i8) cnst1); \ +} +#define DOTP_SB2_SH(...) DOTP_SB2(v8i16, __VA_ARGS__) + +#define DOTP_SB3(RTYPE, mult0, mult1, mult2, cnst0, cnst1, cnst2, \ + out0, out1, out2) \ +{ \ + DOTP_SB2(RTYPE, mult0, mult1, cnst0, cnst1, out0, out1); \ + out2 = (RTYPE) __msa_dotp_s_h((v16i8) mult2, (v16i8) cnst2); \ +} +#define DOTP_SB3_SH(...) DOTP_SB3(v8i16, __VA_ARGS__) + +#define DOTP_SB4(RTYPE, mult0, mult1, mult2, mult3, \ + cnst0, cnst1, cnst2, cnst3, out0, out1, out2, out3) \ +{ \ + DOTP_SB2(RTYPE, mult0, mult1, cnst0, cnst1, out0, out1); \ + DOTP_SB2(RTYPE, mult2, mult3, cnst2, cnst3, out2, out3); \ +} +#define DOTP_SB4_SH(...) DOTP_SB4(v8i16, __VA_ARGS__) + +/* Description : Dot product of halfword vector elements + Arguments : Inputs - mult0, mult1 + cnst0, cnst1 + Outputs - out0, out1 + Return Type - signed word + Details : Signed halfword elements from mult0 are multiplied with + signed halfword elements from cnst0 producing a result + twice the size of input i.e. signed word. + Then this multiplication results of adjacent odd-even elements + are added together and stored to the out vector + (2 signed word results) +*/ +#define DOTP_SH2(RTYPE, mult0, mult1, cnst0, cnst1, out0, out1) \ +{ \ + out0 = (RTYPE) __msa_dotp_s_w((v8i16) mult0, (v8i16) cnst0); \ + out1 = (RTYPE) __msa_dotp_s_w((v8i16) mult1, (v8i16) cnst1); \ +} +#define DOTP_SH2_SW(...) DOTP_SH2(v4i32, __VA_ARGS__) + +#define DOTP_SH4(RTYPE, mult0, mult1, mult2, mult3, \ + cnst0, cnst1, cnst2, cnst3, \ + out0, out1, out2, out3) \ +{ \ + DOTP_SH2(RTYPE, mult0, mult1, cnst0, cnst1, out0, out1); \ + DOTP_SH2(RTYPE, mult2, mult3, cnst2, cnst3, out2, out3); \ +} +#define DOTP_SH4_SW(...) DOTP_SH4(v4i32, __VA_ARGS__) + +/* Description : Dot product & addition of byte vector elements + Arguments : Inputs - mult0, mult1 + cnst0, cnst1 + Outputs - out0, out1 + Return Type - signed halfword + Details : Signed byte elements from mult0 are multiplied with + signed byte elements from cnst0 producing a result + twice the size of input i.e. signed halfword. + Then this multiplication results of adjacent odd-even elements + are added to the out vector + (2 signed halfword results) +*/ +#define DPADD_SB2(RTYPE, mult0, mult1, cnst0, cnst1, out0, out1) \ +{ \ + out0 = (RTYPE) __msa_dpadd_s_h((v8i16) out0, \ + (v16i8) mult0, (v16i8) cnst0); \ + out1 = (RTYPE) __msa_dpadd_s_h((v8i16) out1, \ + (v16i8) mult1, (v16i8) cnst1); \ +} +#define DPADD_SB2_SH(...) DPADD_SB2(v8i16, __VA_ARGS__) + +#define DPADD_SB4(RTYPE, mult0, mult1, mult2, mult3, \ + cnst0, cnst1, cnst2, cnst3, out0, out1, out2, out3) \ +{ \ + DPADD_SB2(RTYPE, mult0, mult1, cnst0, cnst1, out0, out1); \ + DPADD_SB2(RTYPE, mult2, mult3, cnst2, cnst3, out2, out3); \ +} +#define DPADD_SB4_SH(...) DPADD_SB4(v8i16, __VA_ARGS__) + +/* Description : Dot product & addition of byte vector elements + Arguments : Inputs - mult0, mult1 + cnst0, cnst1 + Outputs - out0, out1 + Return Type - unsigned halfword + Details : Unsigned byte elements from mult0 are multiplied with + unsigned byte elements from cnst0 producing a result + twice the size of input i.e. unsigned halfword. + Then this multiplication results of adjacent odd-even elements + are added to the out vector + (2 unsigned halfword results) +*/ +#define DPADD_UB2(RTYPE, mult0, mult1, cnst0, cnst1, out0, out1) \ +{ \ + out0 = (RTYPE) __msa_dpadd_u_h((v8u16) out0, \ + (v16u8) mult0, (v16u8) cnst0); \ + out1 = (RTYPE) __msa_dpadd_u_h((v8u16) out1, \ + (v16u8) mult1, (v16u8) cnst1); \ +} +#define DPADD_UB2_UH(...) DPADD_UB2(v8u16, __VA_ARGS__) + +/* Description : Dot product & addition of halfword vector elements + Arguments : Inputs - mult0, mult1 + cnst0, cnst1 + Outputs - out0, out1 + Return Type - signed word + Details : Signed halfword elements from mult0 are multiplied with + signed halfword elements from cnst0 producing a result + twice the size of input i.e. signed word. + Then this multiplication results of adjacent odd-even elements + are added to the out vector + (2 signed word results) +*/ +#define DPADD_SH2(RTYPE, mult0, mult1, cnst0, cnst1, out0, out1) \ +{ \ + out0 = (RTYPE) __msa_dpadd_s_w((v4i32) out0, \ + (v8i16) mult0, (v8i16) cnst0); \ + out1 = (RTYPE) __msa_dpadd_s_w((v4i32) out1, \ + (v8i16) mult1, (v8i16) cnst1); \ +} +#define DPADD_SH2_SW(...) DPADD_SH2(v4i32, __VA_ARGS__) + +#define DPADD_SH4(RTYPE, mult0, mult1, mult2, mult3, \ + cnst0, cnst1, cnst2, cnst3, out0, out1, out2, out3) \ +{ \ + DPADD_SH2(RTYPE, mult0, mult1, cnst0, cnst1, out0, out1); \ + DPADD_SH2(RTYPE, mult2, mult3, cnst2, cnst3, out2, out3); \ +} +#define DPADD_SH4_SW(...) DPADD_SH4(v4i32, __VA_ARGS__) + +/* Description : Minimum values between unsigned elements of + either vector are copied to the output vector + Arguments : Inputs - in0, in1, min_vec + Outputs - in0, in1, (in place) + Return Type - unsigned halfword + Details : Minimum of unsigned halfword element values from 'in0' and + 'min_value' are written to output vector 'in0' +*/ +#define MIN_UH2(RTYPE, in0, in1, min_vec) \ +{ \ + in0 = (RTYPE) __msa_min_u_h((v8u16) in0, min_vec); \ + in1 = (RTYPE) __msa_min_u_h((v8u16) in1, min_vec); \ +} +#define MIN_UH2_UH(...) MIN_UH2(v8u16, __VA_ARGS__) + +#define MIN_UH4(RTYPE, in0, in1, in2, in3, min_vec) \ +{ \ + MIN_UH2(RTYPE, in0, in1, min_vec); \ + MIN_UH2(RTYPE, in2, in3, min_vec); \ +} +#define MIN_UH4_UH(...) MIN_UH4(v8u16, __VA_ARGS__) + +/* Description : Clips all halfword elements of input vector between min & max + out = ((in) < (min)) ? (min) : (((in) > (max)) ? (max) : (in)) + Arguments : Inputs - in (input vector) + - min (min threshold) + - max (max threshold) + Outputs - out_m (output vector with clipped elements) + Return Type - signed halfword +*/ +#define CLIP_SH(in, min, max) \ +( { \ + v8i16 out_m; \ + \ + out_m = __msa_max_s_h((v8i16) min, (v8i16) in); \ + out_m = __msa_min_s_h((v8i16) max, (v8i16) out_m); \ + out_m; \ +} ) + +/* Description : Clips all signed halfword elements of input vector + between 0 & 255 + Arguments : Inputs - in (input vector) + Outputs - out_m (output vector with clipped elements) + Return Type - signed halfword +*/ +#define CLIP_SH_0_255(in) \ +( { \ + v8i16 max_m = __msa_ldi_h(255); \ + v8i16 out_m; \ + \ + out_m = __msa_maxi_s_h((v8i16) in, 0); \ + out_m = __msa_min_s_h((v8i16) max_m, (v8i16) out_m); \ + out_m; \ +} ) +#define CLIP_SH2_0_255(in0, in1) \ +{ \ + in0 = CLIP_SH_0_255(in0); \ + in1 = CLIP_SH_0_255(in1); \ +} +#define CLIP_SH4_0_255(in0, in1, in2, in3) \ +{ \ + CLIP_SH2_0_255(in0, in1); \ + CLIP_SH2_0_255(in2, in3); \ +} + +/* Description : Clips all signed word elements of input vector + between 0 & 255 + Arguments : Inputs - in (input vector) + Outputs - out_m (output vector with clipped elements) + Return Type - signed word +*/ +#define CLIP_SW_0_255(in) \ +( { \ + v4i32 max_m = __msa_ldi_w(255); \ + v4i32 out_m; \ + \ + out_m = __msa_maxi_s_w((v4i32) in, 0); \ + out_m = __msa_min_s_w((v4i32) max_m, (v4i32) out_m); \ + out_m; \ +} ) + +/* Description : Addition of 4 signed word elements + 4 signed word elements of input vector are added together and + resulted integer sum is returned + Arguments : Inputs - in (signed word vector) + Outputs - sum_m (i32 sum) + Return Type - signed word +*/ +#define HADD_SW_S32(in) \ +( { \ + v2i64 res0_m, res1_m; \ + int32_t sum_m; \ + \ + res0_m = __msa_hadd_s_d((v4i32) in, (v4i32) in); \ + res1_m = __msa_splati_d(res0_m, 1); \ + res0_m = res0_m + res1_m; \ + sum_m = __msa_copy_s_w((v4i32) res0_m, 0); \ + sum_m; \ +} ) + +/* Description : Addition of 8 unsigned halfword elements + 8 unsigned halfword elements of input vector are added + together and resulted integer sum is returned + Arguments : Inputs - in (unsigned halfword vector) + Outputs - sum_m (u32 sum) + Return Type - unsigned word +*/ +#define HADD_UH_U32(in) \ +( { \ + v4u32 res_m; \ + v2u64 res0_m, res1_m; \ + uint32_t sum_m; \ + \ + res_m = __msa_hadd_u_w((v8u16) in, (v8u16) in); \ + res0_m = __msa_hadd_u_d(res_m, res_m); \ + res1_m = (v2u64) __msa_splati_d((v2i64) res0_m, 1); \ + res0_m = res0_m + res1_m; \ + sum_m = __msa_copy_u_w((v4i32) res0_m, 0); \ + sum_m; \ +} ) + +/* Description : Horizontal addition of signed byte vector elements + Arguments : Inputs - in0, in1 + Outputs - out0, out1 + Return Type - as per RTYPE + Details : Each signed odd byte element from 'in0' is added to + even signed byte element from 'in0' (pairwise) and the + halfword result is stored in 'out0' +*/ +#define HADD_SB2(RTYPE, in0, in1, out0, out1) \ +{ \ + out0 = (RTYPE) __msa_hadd_s_h((v16i8) in0, (v16i8) in0); \ + out1 = (RTYPE) __msa_hadd_s_h((v16i8) in1, (v16i8) in1); \ +} +#define HADD_SB2_SH(...) HADD_SB2(v8i16, __VA_ARGS__) + +#define HADD_SB4(RTYPE, in0, in1, in2, in3, out0, out1, out2, out3) \ +{ \ + HADD_SB2(RTYPE, in0, in1, out0, out1); \ + HADD_SB2(RTYPE, in2, in3, out2, out3); \ +} +#define HADD_SB4_UH(...) HADD_SB4(v8u16, __VA_ARGS__) +#define HADD_SB4_SH(...) HADD_SB4(v8i16, __VA_ARGS__) + +/* Description : Horizontal addition of unsigned byte vector elements + Arguments : Inputs - in0, in1 + Outputs - out0, out1 + Return Type - as per RTYPE + Details : Each unsigned odd byte element from 'in0' is added to + even unsigned byte element from 'in0' (pairwise) and the + halfword result is stored in 'out0' +*/ +#define HADD_UB2(RTYPE, in0, in1, out0, out1) \ +{ \ + out0 = (RTYPE) __msa_hadd_u_h((v16u8) in0, (v16u8) in0); \ + out1 = (RTYPE) __msa_hadd_u_h((v16u8) in1, (v16u8) in1); \ +} +#define HADD_UB2_UH(...) HADD_UB2(v8u16, __VA_ARGS__) + +#define HADD_UB3(RTYPE, in0, in1, in2, out0, out1, out2) \ +{ \ + HADD_UB2(RTYPE, in0, in1, out0, out1); \ + out2 = (RTYPE) __msa_hadd_u_h((v16u8) in2, (v16u8) in2); \ +} +#define HADD_UB3_UH(...) HADD_UB3(v8u16, __VA_ARGS__) + +#define HADD_UB4(RTYPE, in0, in1, in2, in3, out0, out1, out2, out3) \ +{ \ + HADD_UB2(RTYPE, in0, in1, out0, out1); \ + HADD_UB2(RTYPE, in2, in3, out2, out3); \ +} +#define HADD_UB4_UB(...) HADD_UB4(v16u8, __VA_ARGS__) +#define HADD_UB4_UH(...) HADD_UB4(v8u16, __VA_ARGS__) +#define HADD_UB4_SH(...) HADD_UB4(v8i16, __VA_ARGS__) + +/* Description : Horizontal subtraction of unsigned byte vector elements + Arguments : Inputs - in0, in1 + Outputs - out0, out1 + Return Type - as per RTYPE + Details : Each unsigned odd byte element from 'in0' is subtracted from + even unsigned byte element from 'in0' (pairwise) and the + halfword result is stored in 'out0' +*/ +#define HSUB_UB2(RTYPE, in0, in1, out0, out1) \ +{ \ + out0 = (RTYPE) __msa_hsub_u_h((v16u8) in0, (v16u8) in0); \ + out1 = (RTYPE) __msa_hsub_u_h((v16u8) in1, (v16u8) in1); \ +} +#define HSUB_UB2_UH(...) HSUB_UB2(v8u16, __VA_ARGS__) +#define HSUB_UB2_SH(...) HSUB_UB2(v8i16, __VA_ARGS__) + +#define HSUB_UB4(RTYPE, in0, in1, in2, in3, out0, out1, out2, out3) \ +{ \ + HSUB_UB2(RTYPE, in0, in1, out0, out1); \ + HSUB_UB2(RTYPE, in2, in3, out2, out3); \ +} +#define HSUB_UB4_UH(...) HSUB_UB4(v8u16, __VA_ARGS__) +#define HSUB_UB4_SH(...) HSUB_UB4(v8i16, __VA_ARGS__) + +/* Description : SAD (Sum of Absolute Difference) + Arguments : Inputs - in0, in1, ref0, ref1 (unsigned byte src & ref) + Outputs - sad_m (halfword vector with sad) + Return Type - unsigned halfword + Details : Absolute difference of all the byte elements from 'in0' with + 'ref0' is calculated and preserved in 'diff0'. From the 16 + unsigned absolute diff values, even-odd pairs are added + together to generate 8 halfword results. +*/ +#define SAD_UB2_UH(in0, in1, ref0, ref1) \ +( { \ + v16u8 diff0_m, diff1_m; \ + v8u16 sad_m = { 0 }; \ + \ + diff0_m = __msa_asub_u_b((v16u8) in0, (v16u8) ref0); \ + diff1_m = __msa_asub_u_b((v16u8) in1, (v16u8) ref1); \ + \ + sad_m += __msa_hadd_u_h((v16u8) diff0_m, (v16u8) diff0_m); \ + sad_m += __msa_hadd_u_h((v16u8) diff1_m, (v16u8) diff1_m); \ + \ + sad_m; \ +} ) + +/* Description : Insert specified word elements from input vectors to 1 + destination vector + Arguments : Inputs - in0, in1, in2, in3 (4 input vectors) + Outputs - out (output vector) + Return Type - as per RTYPE +*/ +#define INSERT_W2(RTYPE, in0, in1, out) \ +{ \ + out = (RTYPE) __msa_insert_w((v4i32) out, 0, in0); \ + out = (RTYPE) __msa_insert_w((v4i32) out, 1, in1); \ +} +#define INSERT_W2_UB(...) INSERT_W2(v16u8, __VA_ARGS__) +#define INSERT_W2_SB(...) INSERT_W2(v16i8, __VA_ARGS__) + +#define INSERT_W4(RTYPE, in0, in1, in2, in3, out) \ +{ \ + out = (RTYPE) __msa_insert_w((v4i32) out, 0, in0); \ + out = (RTYPE) __msa_insert_w((v4i32) out, 1, in1); \ + out = (RTYPE) __msa_insert_w((v4i32) out, 2, in2); \ + out = (RTYPE) __msa_insert_w((v4i32) out, 3, in3); \ +} +#define INSERT_W4_UB(...) INSERT_W4(v16u8, __VA_ARGS__) +#define INSERT_W4_SB(...) INSERT_W4(v16i8, __VA_ARGS__) +#define INSERT_W4_SW(...) INSERT_W4(v4i32, __VA_ARGS__) + +/* Description : Insert specified double word elements from input vectors to 1 + destination vector + Arguments : Inputs - in0, in1 (2 input vectors) + Outputs - out (output vector) + Return Type - as per RTYPE +*/ +#define INSERT_D2(RTYPE, in0, in1, out) \ +{ \ + out = (RTYPE) __msa_insert_d((v2i64) out, 0, in0); \ + out = (RTYPE) __msa_insert_d((v2i64) out, 1, in1); \ +} +#define INSERT_D2_UB(...) INSERT_D2(v16u8, __VA_ARGS__) +#define INSERT_D2_SB(...) INSERT_D2(v16i8, __VA_ARGS__) +#define INSERT_D2_SD(...) INSERT_D2(v2i64, __VA_ARGS__) + +/* Description : Interleave even byte elements from vectors + Arguments : Inputs - in0, in1, in2, in3 + Outputs - out0, out1 + Return Type - as per RTYPE + Details : Even byte elements of 'in0' and even byte + elements of 'in1' are interleaved and copied to 'out0' + Even byte elements of 'in2' and even byte + elements of 'in3' are interleaved and copied to 'out1' +*/ +#define ILVEV_B2(RTYPE, in0, in1, in2, in3, out0, out1) \ +{ \ + out0 = (RTYPE) __msa_ilvev_b((v16i8) in1, (v16i8) in0); \ + out1 = (RTYPE) __msa_ilvev_b((v16i8) in3, (v16i8) in2); \ +} +#define ILVEV_B2_UB(...) ILVEV_B2(v16u8, __VA_ARGS__) +#define ILVEV_B2_SB(...) ILVEV_B2(v16i8, __VA_ARGS__) +#define ILVEV_B2_SH(...) ILVEV_B2(v8i16, __VA_ARGS__) +#define ILVEV_B2_SD(...) ILVEV_B2(v2i64, __VA_ARGS__) + +/* Description : Interleave even halfword elements from vectors + Arguments : Inputs - in0, in1, in2, in3 + Outputs - out0, out1 + Return Type - as per RTYPE + Details : Even halfword elements of 'in0' and even halfword + elements of 'in1' are interleaved and copied to 'out0' + Even halfword elements of 'in2' and even halfword + elements of 'in3' are interleaved and copied to 'out1' +*/ +#define ILVEV_H2(RTYPE, in0, in1, in2, in3, out0, out1) \ +{ \ + out0 = (RTYPE) __msa_ilvev_h((v8i16) in1, (v8i16) in0); \ + out1 = (RTYPE) __msa_ilvev_h((v8i16) in3, (v8i16) in2); \ +} +#define ILVEV_H2_UB(...) ILVEV_H2(v16u8, __VA_ARGS__) +#define ILVEV_H2_SH(...) ILVEV_H2(v8i16, __VA_ARGS__) +#define ILVEV_H2_SW(...) ILVEV_H2(v4i32, __VA_ARGS__) + +/* Description : Interleave even word elements from vectors + Arguments : Inputs - in0, in1, in2, in3 + Outputs - out0, out1 + Return Type - as per RTYPE + Details : Even word elements of 'in0' and even word + elements of 'in1' are interleaved and copied to 'out0' + Even word elements of 'in2' and even word + elements of 'in3' are interleaved and copied to 'out1' +*/ +#define ILVEV_W2(RTYPE, in0, in1, in2, in3, out0, out1) \ +{ \ + out0 = (RTYPE) __msa_ilvev_w((v4i32) in1, (v4i32) in0); \ + out1 = (RTYPE) __msa_ilvev_w((v4i32) in3, (v4i32) in2); \ +} +#define ILVEV_W2_UB(...) ILVEV_W2(v16u8, __VA_ARGS__) +#define ILVEV_W2_SB(...) ILVEV_W2(v16i8, __VA_ARGS__) +#define ILVEV_W2_UH(...) ILVEV_W2(v8u16, __VA_ARGS__) +#define ILVEV_W2_SD(...) ILVEV_W2(v2i64, __VA_ARGS__) + +/* Description : Interleave even double word elements from vectors + Arguments : Inputs - in0, in1, in2, in3 + Outputs - out0, out1 + Return Type - as per RTYPE + Details : Even double word elements of 'in0' and even double word + elements of 'in1' are interleaved and copied to 'out0' + Even double word elements of 'in2' and even double word + elements of 'in3' are interleaved and copied to 'out1' +*/ +#define ILVEV_D2(RTYPE, in0, in1, in2, in3, out0, out1) \ +{ \ + out0 = (RTYPE) __msa_ilvev_d((v2i64) in1, (v2i64) in0); \ + out1 = (RTYPE) __msa_ilvev_d((v2i64) in3, (v2i64) in2); \ +} +#define ILVEV_D2_UB(...) ILVEV_D2(v16u8, __VA_ARGS__) +#define ILVEV_D2_SB(...) ILVEV_D2(v16i8, __VA_ARGS__) +#define ILVEV_D2_SW(...) ILVEV_D2(v4i32, __VA_ARGS__) + +/* Description : Interleave left half of byte elements from vectors + Arguments : Inputs - in0, in1, in2, in3 + Outputs - out0, out1 + Return Type - as per RTYPE + Details : Left half of byte elements of in0 and left half of byte + elements of in1 are interleaved and copied to out0. + Left half of byte elements of in2 and left half of byte + elements of in3 are interleaved and copied to out1. +*/ +#define ILVL_B2(RTYPE, in0, in1, in2, in3, out0, out1) \ +{ \ + out0 = (RTYPE) __msa_ilvl_b((v16i8) in0, (v16i8) in1); \ + out1 = (RTYPE) __msa_ilvl_b((v16i8) in2, (v16i8) in3); \ +} +#define ILVL_B2_UB(...) ILVL_B2(v16u8, __VA_ARGS__) +#define ILVL_B2_SB(...) ILVL_B2(v16i8, __VA_ARGS__) +#define ILVL_B2_UH(...) ILVL_B2(v8u16, __VA_ARGS__) +#define ILVL_B2_SH(...) ILVL_B2(v8i16, __VA_ARGS__) + +#define ILVL_B4(RTYPE, in0, in1, in2, in3, in4, in5, in6, in7, \ + out0, out1, out2, out3) \ +{ \ + ILVL_B2(RTYPE, in0, in1, in2, in3, out0, out1); \ + ILVL_B2(RTYPE, in4, in5, in6, in7, out2, out3); \ +} +#define ILVL_B4_UB(...) ILVL_B4(v16u8, __VA_ARGS__) +#define ILVL_B4_SB(...) ILVL_B4(v16i8, __VA_ARGS__) +#define ILVL_B4_UH(...) ILVL_B4(v8u16, __VA_ARGS__) +#define ILVL_B4_SH(...) ILVL_B4(v8i16, __VA_ARGS__) + +/* Description : Interleave left half of halfword elements from vectors + Arguments : Inputs - in0, in1, in2, in3 + Outputs - out0, out1 + Return Type - as per RTYPE + Details : Left half of halfword elements of in0 and left half of halfword + elements of in1 are interleaved and copied to out0. + Left half of halfword elements of in2 and left half of halfword + elements of in3 are interleaved and copied to out1. +*/ +#define ILVL_H2(RTYPE, in0, in1, in2, in3, out0, out1) \ +{ \ + out0 = (RTYPE) __msa_ilvl_h((v8i16) in0, (v8i16) in1); \ + out1 = (RTYPE) __msa_ilvl_h((v8i16) in2, (v8i16) in3); \ +} +#define ILVL_H2_SH(...) ILVL_H2(v8i16, __VA_ARGS__) +#define ILVL_H2_SW(...) ILVL_H2(v4i32, __VA_ARGS__) + +#define ILVL_H4(RTYPE, in0, in1, in2, in3, in4, in5, in6, in7, \ + out0, out1, out2, out3) \ +{ \ + ILVL_H2(RTYPE, in0, in1, in2, in3, out0, out1); \ + ILVL_H2(RTYPE, in4, in5, in6, in7, out2, out3); \ +} +#define ILVL_H4_SH(...) ILVL_H4(v8i16, __VA_ARGS__) +#define ILVL_H4_SW(...) ILVL_H4(v4i32, __VA_ARGS__) + +/* Description : Interleave left half of word elements from vectors + Arguments : Inputs - in0, in1, in2, in3 + Outputs - out0, out1 + Return Type - as per RTYPE + Details : Left half of word elements of in0 and left half of word + elements of in1 are interleaved and copied to out0. + Left half of word elements of in2 and left half of word + elements of in3 are interleaved and copied to out1. +*/ +#define ILVL_W2(RTYPE, in0, in1, in2, in3, out0, out1) \ +{ \ + out0 = (RTYPE) __msa_ilvl_w((v4i32) in0, (v4i32) in1); \ + out1 = (RTYPE) __msa_ilvl_w((v4i32) in2, (v4i32) in3); \ +} +#define ILVL_W2_UB(...) ILVL_W2(v16u8, __VA_ARGS__) +#define ILVL_W2_SB(...) ILVL_W2(v16i8, __VA_ARGS__) +#define ILVL_W2_SH(...) ILVL_W2(v8i16, __VA_ARGS__) + +/* Description : Interleave right half of byte elements from vectors + Arguments : Inputs - in0, in1, in2, in3, in4, in5, in6, in7 + Outputs - out0, out1, out2, out3 + Return Type - as per RTYPE + Details : Right half of byte elements of in0 and right half of byte + elements of in1 are interleaved and copied to out0. + Right half of byte elements of in2 and right half of byte + elements of in3 are interleaved and copied to out1. + Similar for other pairs +*/ +#define ILVR_B2(RTYPE, in0, in1, in2, in3, out0, out1) \ +{ \ + out0 = (RTYPE) __msa_ilvr_b((v16i8) in0, (v16i8) in1); \ + out1 = (RTYPE) __msa_ilvr_b((v16i8) in2, (v16i8) in3); \ +} +#define ILVR_B2_UB(...) ILVR_B2(v16u8, __VA_ARGS__) +#define ILVR_B2_SB(...) ILVR_B2(v16i8, __VA_ARGS__) +#define ILVR_B2_UH(...) ILVR_B2(v8u16, __VA_ARGS__) +#define ILVR_B2_SH(...) ILVR_B2(v8i16, __VA_ARGS__) +#define ILVR_B2_SW(...) ILVR_B2(v4i32, __VA_ARGS__) + +#define ILVR_B3(RTYPE, in0, in1, in2, in3, in4, in5, out0, out1, out2) \ +{ \ + ILVR_B2(RTYPE, in0, in1, in2, in3, out0, out1); \ + out2 = (RTYPE) __msa_ilvr_b((v16i8) in4, (v16i8) in5); \ +} +#define ILVR_B3_UB(...) ILVR_B3(v16u8, __VA_ARGS__) +#define ILVR_B3_UH(...) ILVR_B3(v8u16, __VA_ARGS__) +#define ILVR_B3_SH(...) ILVR_B3(v8i16, __VA_ARGS__) + +#define ILVR_B4(RTYPE, in0, in1, in2, in3, in4, in5, in6, in7, \ + out0, out1, out2, out3) \ +{ \ + ILVR_B2(RTYPE, in0, in1, in2, in3, out0, out1); \ + ILVR_B2(RTYPE, in4, in5, in6, in7, out2, out3); \ +} +#define ILVR_B4_UB(...) ILVR_B4(v16u8, __VA_ARGS__) +#define ILVR_B4_SB(...) ILVR_B4(v16i8, __VA_ARGS__) +#define ILVR_B4_UH(...) ILVR_B4(v8u16, __VA_ARGS__) +#define ILVR_B4_SH(...) ILVR_B4(v8i16, __VA_ARGS__) +#define ILVR_B4_SW(...) ILVR_B4(v4i32, __VA_ARGS__) + +#define ILVR_B8(RTYPE, in0, in1, in2, in3, in4, in5, in6, in7, \ + in8, in9, in10, in11, in12, in13, in14, in15, \ + out0, out1, out2, out3, out4, out5, out6, out7) \ +{ \ + ILVR_B4(RTYPE, in0, in1, in2, in3, in4, in5, in6, in7, \ + out0, out1, out2, out3); \ + ILVR_B4(RTYPE, in8, in9, in10, in11, in12, in13, in14, in15, \ + out4, out5, out6, out7); \ +} +#define ILVR_B8_UH(...) ILVR_B8(v8u16, __VA_ARGS__) + +/* Description : Interleave right half of halfword elements from vectors + Arguments : Inputs - in0, in1, in2, in3, in4, in5, in6, in7 + Outputs - out0, out1, out2, out3 + Return Type - signed halfword + Details : Right half of halfword elements of in0 and right half of + halfword elements of in1 are interleaved and copied to out0. + Right half of halfword elements of in2 and right half of + halfword elements of in3 are interleaved and copied to out1. + Similar for other pairs +*/ +#define ILVR_H2(RTYPE, in0, in1, in2, in3, out0, out1) \ +{ \ + out0 = (RTYPE) __msa_ilvr_h((v8i16) in0, (v8i16) in1); \ + out1 = (RTYPE) __msa_ilvr_h((v8i16) in2, (v8i16) in3); \ +} +#define ILVR_H2_SH(...) ILVR_H2(v8i16, __VA_ARGS__) +#define ILVR_H2_SW(...) ILVR_H2(v4i32, __VA_ARGS__) + +#define ILVR_H3(RTYPE, in0, in1, in2, in3, in4, in5, out0, out1, out2) \ +{ \ + ILVR_H2(RTYPE, in0, in1, in2, in3, out0, out1); \ + out2 = (RTYPE) __msa_ilvr_h((v8i16) in4, (v8i16) in5); \ +} +#define ILVR_H3_SH(...) ILVR_H3(v8i16, __VA_ARGS__) + +#define ILVR_H4(RTYPE, in0, in1, in2, in3, in4, in5, in6, in7, \ + out0, out1, out2, out3) \ +{ \ + ILVR_H2(RTYPE, in0, in1, in2, in3, out0, out1); \ + ILVR_H2(RTYPE, in4, in5, in6, in7, out2, out3); \ +} +#define ILVR_H4_SH(...) ILVR_H4(v8i16, __VA_ARGS__) +#define ILVR_H4_SW(...) ILVR_H4(v4i32, __VA_ARGS__) + +#define ILVR_W2(RTYPE, in0, in1, in2, in3, out0, out1) \ +{ \ + out0 = (RTYPE) __msa_ilvr_w((v4i32) in0, (v4i32) in1); \ + out1 = (RTYPE) __msa_ilvr_w((v4i32) in2, (v4i32) in3); \ +} +#define ILVR_W2_UB(...) ILVR_W2(v16u8, __VA_ARGS__) +#define ILVR_W2_SB(...) ILVR_W2(v16i8, __VA_ARGS__) +#define ILVR_W2_SH(...) ILVR_W2(v8i16, __VA_ARGS__) + +#define ILVR_W4(RTYPE, in0, in1, in2, in3, in4, in5, in6, in7, \ + out0, out1, out2, out3) \ +{ \ + ILVR_W2(RTYPE, in0, in1, in2, in3, out0, out1); \ + ILVR_W2(RTYPE, in4, in5, in6, in7, out2, out3); \ +} +#define ILVR_W4_SB(...) ILVR_W4(v16i8, __VA_ARGS__) +#define ILVR_W4_UB(...) ILVR_W4(v16u8, __VA_ARGS__) + +/* Description : Interleave right half of double word elements from vectors + Arguments : Inputs - in0, in1, in2, in3, in4, in5, in6, in7 + Outputs - out0, out1, out2, out3 + Return Type - unsigned double word + Details : Right half of double word elements of in0 and right half of + double word elements of in1 are interleaved and copied to out0. + Right half of double word elements of in2 and right half of + double word elements of in3 are interleaved and copied to out1. +*/ +#define ILVR_D2(RTYPE, in0, in1, in2, in3, out0, out1) \ +{ \ + out0 = (RTYPE) __msa_ilvr_d((v2i64) (in0), (v2i64) (in1)); \ + out1 = (RTYPE) __msa_ilvr_d((v2i64) (in2), (v2i64) (in3)); \ +} +#define ILVR_D2_UB(...) ILVR_D2(v16u8, __VA_ARGS__) +#define ILVR_D2_SB(...) ILVR_D2(v16i8, __VA_ARGS__) +#define ILVR_D2_SH(...) ILVR_D2(v8i16, __VA_ARGS__) + +#define ILVR_D3(RTYPE, in0, in1, in2, in3, in4, in5, out0, out1, out2) \ +{ \ + ILVR_D2(RTYPE, in0, in1, in2, in3, out0, out1); \ + out2 = (RTYPE) __msa_ilvr_d((v2i64) (in4), (v2i64) (in5)); \ +} +#define ILVR_D3_SB(...) ILVR_D3(v16i8, __VA_ARGS__) + +#define ILVR_D4(RTYPE, in0, in1, in2, in3, in4, in5, in6, in7, \ + out0, out1, out2, out3) \ +{ \ + ILVR_D2(RTYPE, in0, in1, in2, in3, out0, out1); \ + ILVR_D2(RTYPE, in4, in5, in6, in7, out2, out3); \ +} +#define ILVR_D4_SB(...) ILVR_D4(v16i8, __VA_ARGS__) +#define ILVR_D4_UB(...) ILVR_D4(v16u8, __VA_ARGS__) + +/* Description : Interleave both left and right half of input vectors + Arguments : Inputs - in0, in1 + Outputs - out0, out1 + Return Type - as per RTYPE + Details : Right half of byte elements from 'in0' and 'in1' are + interleaved and stored to 'out0' + Left half of byte elements from 'in0' and 'in1' are + interleaved and stored to 'out1' +*/ +#define ILVRL_B2(RTYPE, in0, in1, out0, out1) \ +{ \ + out0 = (RTYPE) __msa_ilvr_b((v16i8) in0, (v16i8) in1); \ + out1 = (RTYPE) __msa_ilvl_b((v16i8) in0, (v16i8) in1); \ +} +#define ILVRL_B2_UB(...) ILVRL_B2(v16u8, __VA_ARGS__) +#define ILVRL_B2_SB(...) ILVRL_B2(v16i8, __VA_ARGS__) +#define ILVRL_B2_UH(...) ILVRL_B2(v8u16, __VA_ARGS__) +#define ILVRL_B2_SH(...) ILVRL_B2(v8i16, __VA_ARGS__) +#define ILVRL_B2_SW(...) ILVRL_B2(v4i32, __VA_ARGS__) + +#define ILVRL_H2(RTYPE, in0, in1, out0, out1) \ +{ \ + out0 = (RTYPE) __msa_ilvr_h((v8i16) in0, (v8i16) in1); \ + out1 = (RTYPE) __msa_ilvl_h((v8i16) in0, (v8i16) in1); \ +} +#define ILVRL_H2_SB(...) ILVRL_H2(v16i8, __VA_ARGS__) +#define ILVRL_H2_SH(...) ILVRL_H2(v8i16, __VA_ARGS__) +#define ILVRL_H2_SW(...) ILVRL_H2(v4i32, __VA_ARGS__) + +#define ILVRL_W2(RTYPE, in0, in1, out0, out1) \ +{ \ + out0 = (RTYPE) __msa_ilvr_w((v4i32) in0, (v4i32) in1); \ + out1 = (RTYPE) __msa_ilvl_w((v4i32) in0, (v4i32) in1); \ +} +#define ILVRL_W2_UB(...) ILVRL_W2(v16u8, __VA_ARGS__) +#define ILVRL_W2_SH(...) ILVRL_W2(v8i16, __VA_ARGS__) +#define ILVRL_W2_SW(...) ILVRL_W2(v4i32, __VA_ARGS__) + +/* Description : Maximum values between signed elements of vector and + 5-bit signed immediate value are copied to the output vector + Arguments : Inputs - in0, in1, in2, in3, max_val + Outputs - in0, in1, in2, in3 (in place) + Return Type - unsigned halfword + Details : Maximum of signed halfword element values from 'in0' and + 'max_val' are written to output vector 'in0' +*/ +#define MAXI_SH2(RTYPE, in0, in1, max_val) \ +{ \ + in0 = (RTYPE) __msa_maxi_s_h((v8i16) in0, (max_val)); \ + in1 = (RTYPE) __msa_maxi_s_h((v8i16) in1, (max_val)); \ +} +#define MAXI_SH2_UH(...) MAXI_SH2(v8u16, __VA_ARGS__) +#define MAXI_SH2_SH(...) MAXI_SH2(v8i16, __VA_ARGS__) + +#define MAXI_SH4(RTYPE, in0, in1, in2, in3, max_val) \ +{ \ + MAXI_SH2(RTYPE, in0, in1, max_val); \ + MAXI_SH2(RTYPE, in2, in3, max_val); \ +} +#define MAXI_SH4_UH(...) MAXI_SH4(v8u16, __VA_ARGS__) + +/* Description : Saturate the halfword element values to the max + unsigned value of (sat_val+1 bits) + The element data width remains unchanged + Arguments : Inputs - in0, in1, in2, in3, sat_val + Outputs - in0, in1, in2, in3 (in place) + Return Type - unsigned halfword + Details : Each unsigned halfword element from 'in0' is saturated to the + value generated with (sat_val+1) bit range + Results are in placed to original vectors +*/ +#define SAT_UH2(RTYPE, in0, in1, sat_val) \ +{ \ + in0 = (RTYPE) __msa_sat_u_h((v8u16) in0, sat_val); \ + in1 = (RTYPE) __msa_sat_u_h((v8u16) in1, sat_val); \ +} +#define SAT_UH2_UH(...) SAT_UH2(v8u16, __VA_ARGS__) +#define SAT_UH2_SH(...) SAT_UH2(v8i16, __VA_ARGS__) + +#define SAT_UH4(RTYPE, in0, in1, in2, in3, sat_val) \ +{ \ + SAT_UH2(RTYPE, in0, in1, sat_val); \ + SAT_UH2(RTYPE, in2, in3, sat_val) \ +} +#define SAT_UH4_UH(...) SAT_UH4(v8u16, __VA_ARGS__) + +/* Description : Saturate the halfword element values to the max + unsigned value of (sat_val+1 bits) + The element data width remains unchanged + Arguments : Inputs - in0, in1, in2, in3, sat_val + Outputs - in0, in1, in2, in3 (in place) + Return Type - unsigned halfword + Details : Each unsigned halfword element from 'in0' is saturated to the + value generated with (sat_val+1) bit range + Results are in placed to original vectors +*/ +#define SAT_SH2(RTYPE, in0, in1, sat_val) \ +{ \ + in0 = (RTYPE) __msa_sat_s_h((v8i16) in0, sat_val); \ + in1 = (RTYPE) __msa_sat_s_h((v8i16) in1, sat_val); \ +} +#define SAT_SH2_SH(...) SAT_SH2(v8i16, __VA_ARGS__) + +#define SAT_SH3(RTYPE, in0, in1, in2, sat_val) \ +{ \ + SAT_SH2(RTYPE, in0, in1, sat_val) \ + in2 = (RTYPE) __msa_sat_s_h((v8i16) in2, sat_val); \ +} +#define SAT_SH3_SH(...) SAT_SH3(v8i16, __VA_ARGS__) + +#define SAT_SH4(RTYPE, in0, in1, in2, in3, sat_val) \ +{ \ + SAT_SH2(RTYPE, in0, in1, sat_val); \ + SAT_SH2(RTYPE, in2, in3, sat_val); \ +} +#define SAT_SH4_SH(...) SAT_SH4(v8i16, __VA_ARGS__) + +/* Description : Saturate the word element values to the max + unsigned value of (sat_val+1 bits) + The element data width remains unchanged + Arguments : Inputs - in0, in1, in2, in3, sat_val + Outputs - in0, in1, in2, in3 (in place) + Return Type - unsigned word + Details : Each unsigned word element from 'in0' is saturated to the + value generated with (sat_val+1) bit range + Results are in placed to original vectors +*/ +#define SAT_SW2(RTYPE, in0, in1, sat_val) \ +{ \ + in0 = (RTYPE) __msa_sat_s_w((v4i32) in0, sat_val); \ + in1 = (RTYPE) __msa_sat_s_w((v4i32) in1, sat_val); \ +} +#define SAT_SW2_SW(...) SAT_SW2(v4i32, __VA_ARGS__) + +#define SAT_SW4(RTYPE, in0, in1, in2, in3, sat_val) \ +{ \ + SAT_SW2(RTYPE, in0, in1, sat_val); \ + SAT_SW2(RTYPE, in2, in3, sat_val); \ +} +#define SAT_SW4_SW(...) SAT_SW4(v4i32, __VA_ARGS__) + +/* Description : Indexed halfword element values are replicated to all + elements in output vector + Arguments : Inputs - in, idx0, idx1 + Outputs - out0, out1 + Return Type - as per RTYPE + Details : 'idx0' element value from 'in' vector is replicated to all + elements in 'out0' vector + Valid index range for halfword operation is 0-7 +*/ +#define SPLATI_H2(RTYPE, in, idx0, idx1, out0, out1) \ +{ \ + out0 = (RTYPE) __msa_splati_h((v8i16) in, idx0); \ + out1 = (RTYPE) __msa_splati_h((v8i16) in, idx1); \ +} +#define SPLATI_H2_SB(...) SPLATI_H2(v16i8, __VA_ARGS__) +#define SPLATI_H2_SH(...) SPLATI_H2(v8i16, __VA_ARGS__) + +#define SPLATI_H3(RTYPE, in, idx0, idx1, idx2, \ + out0, out1, out2) \ +{ \ + SPLATI_H2(RTYPE, in, idx0, idx1, out0, out1); \ + out2 = (RTYPE) __msa_splati_h((v8i16) in, idx2); \ +} +#define SPLATI_H3_SB(...) SPLATI_H3(v16i8, __VA_ARGS__) +#define SPLATI_H3_SH(...) SPLATI_H3(v8i16, __VA_ARGS__) + +#define SPLATI_H4(RTYPE, in, idx0, idx1, idx2, idx3, \ + out0, out1, out2, out3) \ +{ \ + SPLATI_H2(RTYPE, in, idx0, idx1, out0, out1); \ + SPLATI_H2(RTYPE, in, idx2, idx3, out2, out3); \ +} +#define SPLATI_H4_SB(...) SPLATI_H4(v16i8, __VA_ARGS__) +#define SPLATI_H4_SH(...) SPLATI_H4(v8i16, __VA_ARGS__) + +/* Description : Indexed word element values are replicated to all + elements in output vector + Arguments : Inputs - in, stidx + Outputs - out0, out1 + Return Type - as per RTYPE + Details : 'stidx' element value from 'in' vector is replicated to all + elements in 'out0' vector + 'stidx + 1' element value from 'in' vector is replicated to all + elements in 'out1' vector + Valid index range for halfword operation is 0-3 +*/ +#define SPLATI_W2(RTYPE, in, stidx, out0, out1) \ +{ \ + out0 = (RTYPE) __msa_splati_w((v4i32) in, stidx); \ + out1 = (RTYPE) __msa_splati_w((v4i32) in, (stidx+1)); \ +} +#define SPLATI_W2_SH(...) SPLATI_W2(v8i16, __VA_ARGS__) +#define SPLATI_W2_SW(...) SPLATI_W2(v4i32, __VA_ARGS__) + +#define SPLATI_W4(RTYPE, in, out0, out1, out2, out3) \ +{ \ + SPLATI_W2(RTYPE, in, 0, out0, out1); \ + SPLATI_W2(RTYPE, in, 2, out2, out3); \ +} +#define SPLATI_W4_SH(...) SPLATI_W4(v8i16, __VA_ARGS__) +#define SPLATI_W4_SW(...) SPLATI_W4(v4i32, __VA_ARGS__) + +/* Description : Pack even byte elements of vector pairs + Arguments : Inputs - in0, in1, in2, in3 + Outputs - out0, out1 + Return Type - as per RTYPE + Details : Even byte elements of in0 are copied to the left half of + out0 & even byte elements of in1 are copied to the right + half of out0. + Even byte elements of in2 are copied to the left half of + out1 & even byte elements of in3 are copied to the right + half of out1. +*/ +#define PCKEV_B2(RTYPE, in0, in1, in2, in3, out0, out1) \ +{ \ + out0 = (RTYPE) __msa_pckev_b((v16i8) in0, (v16i8) in1); \ + out1 = (RTYPE) __msa_pckev_b((v16i8) in2, (v16i8) in3); \ +} +#define PCKEV_B2_SB(...) PCKEV_B2(v16i8, __VA_ARGS__) +#define PCKEV_B2_UB(...) PCKEV_B2(v16u8, __VA_ARGS__) +#define PCKEV_B2_SH(...) PCKEV_B2(v8i16, __VA_ARGS__) +#define PCKEV_B2_SW(...) PCKEV_B2(v4i32, __VA_ARGS__) + +#define PCKEV_B3(RTYPE, in0, in1, in2, in3, in4, in5, out0, out1, out2) \ +{ \ + PCKEV_B2(RTYPE, in0, in1, in2, in3, out0, out1); \ + out2 = (RTYPE) __msa_pckev_b((v16i8) in4, (v16i8) in5); \ +} +#define PCKEV_B3_UB(...) PCKEV_B3(v16u8, __VA_ARGS__) +#define PCKEV_B3_SB(...) PCKEV_B3(v16i8, __VA_ARGS__) + +#define PCKEV_B4(RTYPE, in0, in1, in2, in3, in4, in5, in6, in7, \ + out0, out1, out2, out3) \ +{ \ + PCKEV_B2(RTYPE, in0, in1, in2, in3, out0, out1); \ + PCKEV_B2(RTYPE, in4, in5, in6, in7, out2, out3); \ +} +#define PCKEV_B4_SB(...) PCKEV_B4(v16i8, __VA_ARGS__) +#define PCKEV_B4_UB(...) PCKEV_B4(v16u8, __VA_ARGS__) +#define PCKEV_B4_SH(...) PCKEV_B4(v8i16, __VA_ARGS__) +#define PCKEV_B4_SW(...) PCKEV_B4(v4i32, __VA_ARGS__) + +/* Description : Pack even halfword elements of vector pairs + Arguments : Inputs - in0, in1, in2, in3 + Outputs - out0, out1 + Return Type - as per RTYPE + Details : Even halfword elements of in0 are copied to the left half of + out0 & even halfword elements of in1 are copied to the right + half of out0. + Even halfword elements of in2 are copied to the left half of + out1 & even halfword elements of in3 are copied to the right + half of out1. +*/ +#define PCKEV_H2(RTYPE, in0, in1, in2, in3, out0, out1) \ +{ \ + out0 = (RTYPE) __msa_pckev_h((v8i16) in0, (v8i16) in1); \ + out1 = (RTYPE) __msa_pckev_h((v8i16) in2, (v8i16) in3); \ +} +#define PCKEV_H2_SH(...) PCKEV_H2(v8i16, __VA_ARGS__) +#define PCKEV_H2_SW(...) PCKEV_H2(v4i32, __VA_ARGS__) + +#define PCKEV_H4(RTYPE, in0, in1, in2, in3, in4, in5, in6, in7, \ + out0, out1, out2, out3) \ +{ \ + PCKEV_H2(RTYPE, in0, in1, in2, in3, out0, out1); \ + PCKEV_H2(RTYPE, in4, in5, in6, in7, out2, out3); \ +} +#define PCKEV_H4_SH(...) PCKEV_H4(v8i16, __VA_ARGS__) +#define PCKEV_H4_SW(...) PCKEV_H4(v4i32, __VA_ARGS__) + +/* Description : Pack even double word elements of vector pairs + Arguments : Inputs - in0, in1, in2, in3 + Outputs - out0, out1 + Return Type - unsigned byte + Details : Even double elements of in0 are copied to the left half of + out0 & even double elements of in1 are copied to the right + half of out0. + Even double elements of in2 are copied to the left half of + out1 & even double elements of in3 are copied to the right + half of out1. +*/ +#define PCKEV_D2(RTYPE, in0, in1, in2, in3, out0, out1) \ +{ \ + out0 = (RTYPE) __msa_pckev_d((v2i64) in0, (v2i64) in1); \ + out1 = (RTYPE) __msa_pckev_d((v2i64) in2, (v2i64) in3); \ +} +#define PCKEV_D2_UB(...) PCKEV_D2(v16u8, __VA_ARGS__) +#define PCKEV_D2_SB(...) PCKEV_D2(v16i8, __VA_ARGS__) +#define PCKEV_D2_SH(...) PCKEV_D2(v8i16, __VA_ARGS__) + +#define PCKEV_D4(RTYPE, in0, in1, in2, in3, in4, in5, in6, in7, \ + out0, out1, out2, out3) \ +{ \ + PCKEV_D2(RTYPE, in0, in1, in2, in3, out0, out1); \ + PCKEV_D2(RTYPE, in4, in5, in6, in7, out2, out3); \ +} +#define PCKEV_D4_UB(...) PCKEV_D4(v16u8, __VA_ARGS__) + +/* Description : Pack odd double word elements of vector pairs + Arguments : Inputs - in0, in1 + Outputs - out0, out1 + Return Type - as per RTYPE + Details : As operation is on same input 'in0' vector, index 1 double word + element is overwritten to index 0 and result is written to out0 + As operation is on same input 'in1' vector, index 1 double word + element is overwritten to index 0 and result is written to out1 +*/ +#define PCKOD_D2(RTYPE, in0, in1, in2, in3, out0, out1) \ +{ \ + out0 = (RTYPE) __msa_pckod_d((v2i64) in0, (v2i64) in1); \ + out1 = (RTYPE) __msa_pckod_d((v2i64) in2, (v2i64) in3); \ +} +#define PCKOD_D2_UB(...) PCKOD_D2(v16u8, __VA_ARGS__) +#define PCKOD_D2_SH(...) PCKOD_D2(v8i16, __VA_ARGS__) +#define PCKOD_D2_SD(...) PCKOD_D2(v2i64, __VA_ARGS__) + +/* Description : Each byte element is logically xor'ed with immediate 128 + Arguments : Inputs - in0, in1 + Outputs - in0, in1 (in-place) + Return Type - as per RTYPE + Details : Each unsigned byte element from input vector 'in0' is + logically xor'ed with 128 and result is in-place stored in + 'in0' vector + Each unsigned byte element from input vector 'in1' is + logically xor'ed with 128 and result is in-place stored in + 'in1' vector + Similar for other pairs +*/ +#define XORI_B2_128(RTYPE, in0, in1) \ +{ \ + in0 = (RTYPE) __msa_xori_b((v16u8) in0, 128); \ + in1 = (RTYPE) __msa_xori_b((v16u8) in1, 128); \ +} +#define XORI_B2_128_UB(...) XORI_B2_128(v16u8, __VA_ARGS__) +#define XORI_B2_128_SB(...) XORI_B2_128(v16i8, __VA_ARGS__) +#define XORI_B2_128_SH(...) XORI_B2_128(v8i16, __VA_ARGS__) + +#define XORI_B3_128(RTYPE, in0, in1, in2) \ +{ \ + XORI_B2_128(RTYPE, in0, in1); \ + in2 = (RTYPE) __msa_xori_b((v16u8) in2, 128); \ +} +#define XORI_B3_128_SB(...) XORI_B3_128(v16i8, __VA_ARGS__) + +#define XORI_B4_128(RTYPE, in0, in1, in2, in3) \ +{ \ + XORI_B2_128(RTYPE, in0, in1); \ + XORI_B2_128(RTYPE, in2, in3); \ +} +#define XORI_B4_128_UB(...) XORI_B4_128(v16u8, __VA_ARGS__) +#define XORI_B4_128_SB(...) XORI_B4_128(v16i8, __VA_ARGS__) +#define XORI_B4_128_SH(...) XORI_B4_128(v8i16, __VA_ARGS__) + +#define XORI_B5_128(RTYPE, in0, in1, in2, in3, in4) \ +{ \ + XORI_B3_128(RTYPE, in0, in1, in2); \ + XORI_B2_128(RTYPE, in3, in4); \ +} +#define XORI_B5_128_SB(...) XORI_B5_128(v16i8, __VA_ARGS__) + +#define XORI_B6_128(RTYPE, in0, in1, in2, in3, in4, in5) \ +{ \ + XORI_B4_128(RTYPE, in0, in1, in2, in3); \ + XORI_B2_128(RTYPE, in4, in5); \ +} +#define XORI_B6_128_SB(...) XORI_B6_128(v16i8, __VA_ARGS__) + +#define XORI_B7_128(RTYPE, in0, in1, in2, in3, in4, in5, in6) \ +{ \ + XORI_B4_128(RTYPE, in0, in1, in2, in3); \ + XORI_B3_128(RTYPE, in4, in5, in6); \ +} +#define XORI_B7_128_SB(...) XORI_B7_128(v16i8, __VA_ARGS__) + +#define XORI_B8_128(RTYPE, in0, in1, in2, in3, in4, in5, in6, in7) \ +{ \ + XORI_B4_128(RTYPE, in0, in1, in2, in3); \ + XORI_B4_128(RTYPE, in4, in5, in6, in7); \ +} +#define XORI_B8_128_SB(...) XORI_B8_128(v16i8, __VA_ARGS__) + +/* Description : Addition of signed halfword elements and signed saturation + Arguments : Inputs - in0, in1, in2, in3 + Outputs - out0, out1 + Return Type - as per RTYPE + Details : Signed halfword elements from 'in0' are added to signed + halfword elements of 'in1'. The result is then signed saturated + between -32768 to +32767 (as per halfword data type) + Similar for other pairs +*/ +#define ADDS_SH2(RTYPE, in0, in1, in2, in3, out0, out1) \ +{ \ + out0 = (RTYPE) __msa_adds_s_h((v8i16) in0, (v8i16) in1); \ + out1 = (RTYPE) __msa_adds_s_h((v8i16) in2, (v8i16) in3); \ +} +#define ADDS_SH2_SH(...) ADDS_SH2(v8i16, __VA_ARGS__) + +#define ADDS_SH4(RTYPE, in0, in1, in2, in3, in4, in5, in6, in7, \ + out0, out1, out2, out3) \ +{ \ + ADDS_SH2(RTYPE, in0, in1, in2, in3, out0, out1); \ + ADDS_SH2(RTYPE, in4, in5, in6, in7, out2, out3); \ +} +#define ADDS_SH4_UH(...) ADDS_SH4(v8u16, __VA_ARGS__) +#define ADDS_SH4_SH(...) ADDS_SH4(v8i16, __VA_ARGS__) + +/* Description : Shift left all elements of vector (generic for all data types) + Arguments : Inputs - in0, in1, in2, in3, shift + Outputs - in0, in1, in2, in3 (in place) + Return Type - as per input vector RTYPE + Details : Each element of vector 'in0' is left shifted by 'shift' and + result is in place written to 'in0' + Similar for other pairs +*/ +#define SLLI_4V(in0, in1, in2, in3, shift) \ +{ \ + in0 = in0 << shift; \ + in1 = in1 << shift; \ + in2 = in2 << shift; \ + in3 = in3 << shift; \ +} + +/* Description : Arithmetic shift right all elements of vector + (generic for all data types) + Arguments : Inputs - in0, in1, in2, in3, shift + Outputs - in0, in1, in2, in3 (in place) + Return Type - as per input vector RTYPE + Details : Each element of vector 'in0' is right shifted by 'shift' and + result is in place written to 'in0' + Here, 'shift' is GP variable passed in + Similar for other pairs +*/ +#define SRA_4V(in0, in1, in2, in3, shift) \ +{ \ + in0 = in0 >> shift; \ + in1 = in1 >> shift; \ + in2 = in2 >> shift; \ + in3 = in3 >> shift; \ +} + +/* Description : Shift right logical all halfword elements of vector + Arguments : Inputs - in0, in1, in2, in3, shift + Outputs - in0, in1, in2, in3 (in place) + Return Type - unsigned halfword + Details : Each element of vector 'in0' is shifted right logical by + number of bits respective element holds in vector 'shift' and + result is in place written to 'in0' + Here, 'shift' is a vector passed in + Similar for other pairs +*/ +#define SRL_H4(RTYPE, in0, in1, in2, in3, shift) \ +{ \ + in0 = (RTYPE) __msa_srl_h((v8i16) in0, (v8i16) shift); \ + in1 = (RTYPE) __msa_srl_h((v8i16) in1, (v8i16) shift); \ + in2 = (RTYPE) __msa_srl_h((v8i16) in2, (v8i16) shift); \ + in3 = (RTYPE) __msa_srl_h((v8i16) in3, (v8i16) shift); \ +} +#define SRL_H4_UH(...) SRL_H4(v8u16, __VA_ARGS__) + +/* Description : Shift right arithmetic rounded halfwords + Arguments : Inputs - in0, in1, shift + Outputs - in0, in1, (in place) + Return Type - unsigned halfword + Details : Each element of vector 'in0' is shifted right arithmetic by + number of bits respective element holds in vector 'shift'. + The last discarded bit is added to shifted value for rounding + and the result is in place written to 'in0' + Here, 'shift' is a vector passed in + Similar for other pairs +*/ +#define SRAR_H2(RTYPE, in0, in1, shift) \ +{ \ + in0 = (RTYPE) __msa_srar_h((v8i16) in0, (v8i16) shift); \ + in1 = (RTYPE) __msa_srar_h((v8i16) in1, (v8i16) shift); \ +} +#define SRAR_H2_UH(...) SRAR_H2(v8u16, __VA_ARGS__) +#define SRAR_H2_SH(...) SRAR_H2(v8i16, __VA_ARGS__) + +#define SRAR_H3(RTYPE, in0, in1, in2, shift) \ +{ \ + SRAR_H2(RTYPE, in0, in1, shift) \ + in2 = (RTYPE) __msa_srar_h((v8i16) in2, (v8i16) shift); \ +} +#define SRAR_H3_SH(...) SRAR_H3(v8i16, __VA_ARGS__) + +#define SRAR_H4(RTYPE, in0, in1, in2, in3, shift) \ +{ \ + SRAR_H2(RTYPE, in0, in1, shift) \ + SRAR_H2(RTYPE, in2, in3, shift) \ +} +#define SRAR_H4_UH(...) SRAR_H4(v8u16, __VA_ARGS__) +#define SRAR_H4_SH(...) SRAR_H4(v8i16, __VA_ARGS__) + +/* Description : Shift right arithmetic rounded words + Arguments : Inputs - in0, in1, shift + Outputs - in0, in1, (in place) + Return Type - as per RTYPE + Details : Each element of vector 'in0' is shifted right arithmetic by + number of bits respective element holds in vector 'shift'. + The last discarded bit is added to shifted value for rounding + and the result is in place written to 'in0' + Here, 'shift' is a vector passed in + Similar for other pairs +*/ +#define SRAR_W2(RTYPE, in0, in1, shift) \ +{ \ + in0 = (RTYPE) __msa_srar_w((v4i32) in0, (v4i32) shift); \ + in1 = (RTYPE) __msa_srar_w((v4i32) in1, (v4i32) shift); \ +} +#define SRAR_W2_SW(...) SRAR_W2(v4i32, __VA_ARGS__) + +#define SRAR_W4(RTYPE, in0, in1, in2, in3, shift) \ +{ \ + SRAR_W2(RTYPE, in0, in1, shift) \ + SRAR_W2(RTYPE, in2, in3, shift) \ +} +#define SRAR_W4_SW(...) SRAR_W4(v4i32, __VA_ARGS__) + +/* Description : Shift right arithmetic rounded (immediate) + Arguments : Inputs - in0, in1, in2, in3, shift + Outputs - in0, in1, in2, in3 (in place) + Return Type - as per RTYPE + Details : Each element of vector 'in0' is shifted right arithmetic by + value in 'shift'. + The last discarded bit is added to shifted value for rounding + and the result is in place written to 'in0' + Similar for other pairs +*/ +#define SRARI_H2(RTYPE, in0, in1, shift) \ +{ \ + in0 = (RTYPE) __msa_srari_h((v8i16) in0, shift); \ + in1 = (RTYPE) __msa_srari_h((v8i16) in1, shift); \ +} +#define SRARI_H2_UH(...) SRARI_H2(v8u16, __VA_ARGS__) +#define SRARI_H2_SH(...) SRARI_H2(v8i16, __VA_ARGS__) + +#define SRARI_H4(RTYPE, in0, in1, in2, in3, shift) \ +{ \ + SRARI_H2(RTYPE, in0, in1, shift); \ + SRARI_H2(RTYPE, in2, in3, shift); \ +} +#define SRARI_H4_UH(...) SRARI_H4(v8u16, __VA_ARGS__) +#define SRARI_H4_SH(...) SRARI_H4(v8i16, __VA_ARGS__) + +/* Description : Shift right arithmetic rounded (immediate) + Arguments : Inputs - in0, in1, shift + Outputs - in0, in1 (in place) + Return Type - as per RTYPE + Details : Each element of vector 'in0' is shifted right arithmetic by + value in 'shift'. + The last discarded bit is added to shifted value for rounding + and the result is in place written to 'in0' + Similar for other pairs +*/ +#define SRARI_W2(RTYPE, in0, in1, shift) \ +{ \ + in0 = (RTYPE) __msa_srari_w((v4i32) in0, shift); \ + in1 = (RTYPE) __msa_srari_w((v4i32) in1, shift); \ +} +#define SRARI_W2_SW(...) SRARI_W2(v4i32, __VA_ARGS__) + +#define SRARI_W4(RTYPE, in0, in1, in2, in3, shift) \ +{ \ + SRARI_W2(RTYPE, in0, in1, shift); \ + SRARI_W2(RTYPE, in2, in3, shift); \ +} +#define SRARI_W4_SH(...) SRARI_W4(v8i16, __VA_ARGS__) +#define SRARI_W4_SW(...) SRARI_W4(v4i32, __VA_ARGS__) + +/* Description : Multiplication of pairs of vectors + Arguments : Inputs - in0, in1, in2, in3 + Outputs - out0, out1 + Details : Each element from 'in0' is multiplied with elements from 'in1' + and result is written to 'out0' + Similar for other pairs +*/ +#define MUL2(in0, in1, in2, in3, out0, out1) \ +{ \ + out0 = in0 * in1; \ + out1 = in2 * in3; \ +} +#define MUL4(in0, in1, in2, in3, in4, in5, in6, in7, out0, out1, out2, out3) \ +{ \ + MUL2(in0, in1, in2, in3, out0, out1); \ + MUL2(in4, in5, in6, in7, out2, out3); \ +} + +/* Description : Addition of 2 pairs of vectors + Arguments : Inputs - in0, in1, in2, in3 + Outputs - out0, out1 + Details : Each element from 2 pairs vectors is added and 2 results are + produced +*/ +#define ADD2(in0, in1, in2, in3, out0, out1) \ +{ \ + out0 = in0 + in1; \ + out1 = in2 + in3; \ +} +#define ADD4(in0, in1, in2, in3, in4, in5, in6, in7, out0, out1, out2, out3) \ +{ \ + ADD2(in0, in1, in2, in3, out0, out1); \ + ADD2(in4, in5, in6, in7, out2, out3); \ +} + +/* Description : Subtraction of 2 pairs of vectors + Arguments : Inputs - in0, in1, in2, in3 + Outputs - out0, out1 + Details : Each element from 2 pairs vectors is subtracted and 2 results + are produced +*/ +#define SUB2(in0, in1, in2, in3, out0, out1) \ +{ \ + out0 = in0 - in1; \ + out1 = in2 - in3; \ +} +#define SUB4(in0, in1, in2, in3, in4, in5, in6, in7, out0, out1, out2, out3) \ +{ \ + out0 = in0 - in1; \ + out1 = in2 - in3; \ + out2 = in4 - in5; \ + out3 = in6 - in7; \ +} + +/* Description : Sign extend halfword elements from right half of the vector + Arguments : Inputs - in (input halfword vector) + Outputs - out (sign extended word vectors) + Return Type - signed word + Details : Sign bit of halfword elements from input vector 'in' is + extracted and interleaved with same vector 'in0' to generate + 4 word elements keeping sign intact +*/ +#define UNPCK_R_SH_SW(in, out) \ +{ \ + v8i16 sign_m; \ + \ + sign_m = __msa_clti_s_h((v8i16) in, 0); \ + out = (v4i32) __msa_ilvr_h(sign_m, (v8i16) in); \ +} + +/* Description : Sign extend byte elements from input vector and return + halfword results in pair of vectors + Arguments : Inputs - in (1 input byte vector) + Outputs - out0, out1 (sign extended 2 halfword vectors) + Return Type - signed halfword + Details : Sign bit of byte elements from input vector 'in' is + extracted and interleaved right with same vector 'in0' to + generate 8 signed halfword elements in 'out0' + Then interleaved left with same vector 'in0' to + generate 8 signed halfword elements in 'out1' +*/ +#define UNPCK_SB_SH(in, out0, out1) \ +{ \ + v16i8 tmp_m; \ + \ + tmp_m = __msa_clti_s_b((v16i8) in, 0); \ + ILVRL_B2_SH(tmp_m, in, out0, out1); \ +} + +/* Description : Zero extend unsigned byte elements to halfword elements + Arguments : Inputs - in (1 input unsigned byte vector) + Outputs - out0, out1 (unsigned 2 halfword vectors) + Return Type - signed halfword + Details : Zero extended right half of vector is returned in 'out0' + Zero extended left half of vector is returned in 'out1' +*/ +#define UNPCK_UB_SH(in, out0, out1) \ +{ \ + v16i8 zero_m = { 0 }; \ + \ + ILVRL_B2_SH(zero_m, in, out0, out1); \ +} + +/* Description : Sign extend halfword elements from input vector and return + result in pair of vectors + Arguments : Inputs - in (1 input halfword vector) + Outputs - out0, out1 (sign extended 2 word vectors) + Return Type - signed word + Details : Sign bit of halfword elements from input vector 'in' is + extracted and interleaved right with same vector 'in0' to + generate 4 signed word elements in 'out0' + Then interleaved left with same vector 'in0' to + generate 4 signed word elements in 'out1' +*/ +#define UNPCK_SH_SW(in, out0, out1) \ +{ \ + v8i16 tmp_m; \ + \ + tmp_m = __msa_clti_s_h((v8i16) in, 0); \ + ILVRL_H2_SW(tmp_m, in, out0, out1); \ +} + +/* Description : Swap two variables + Arguments : Inputs - in0, in1 + Outputs - in0, in1 (in-place) + Details : Swapping of two input variables using xor +*/ +#define SWAP(in0, in1) \ +{ \ + in0 = in0 ^ in1; \ + in1 = in0 ^ in1; \ + in0 = in0 ^ in1; \ +} + +/* Description : Butterfly of 4 input vectors + Arguments : Inputs - in0, in1, in2, in3 + Outputs - out0, out1, out2, out3 + Details : Butterfly operation +*/ +#define BUTTERFLY_4(in0, in1, in2, in3, out0, out1, out2, out3) \ +{ \ + out0 = in0 + in3; \ + out1 = in1 + in2; \ + \ + out2 = in1 - in2; \ + out3 = in0 - in3; \ +} + +/* Description : Butterfly of 8 input vectors + Arguments : Inputs - in0 ... in7 + Outputs - out0 .. out7 + Details : Butterfly operation +*/ +#define BUTTERFLY_8(in0, in1, in2, in3, in4, in5, in6, in7, \ + out0, out1, out2, out3, out4, out5, out6, out7) \ +{ \ + out0 = in0 + in7; \ + out1 = in1 + in6; \ + out2 = in2 + in5; \ + out3 = in3 + in4; \ + \ + out4 = in3 - in4; \ + out5 = in2 - in5; \ + out6 = in1 - in6; \ + out7 = in0 - in7; \ +} + +/* Description : Butterfly of 16 input vectors + Arguments : Inputs - in0 ... in15 + Outputs - out0 .. out15 + Details : Butterfly operation +*/ +#define BUTTERFLY_16(in0, in1, in2, in3, in4, in5, in6, in7, \ + in8, in9, in10, in11, in12, in13, in14, in15, \ + out0, out1, out2, out3, out4, out5, out6, out7, \ + out8, out9, out10, out11, out12, out13, out14, out15) \ +{ \ + out0 = in0 + in15; \ + out1 = in1 + in14; \ + out2 = in2 + in13; \ + out3 = in3 + in12; \ + out4 = in4 + in11; \ + out5 = in5 + in10; \ + out6 = in6 + in9; \ + out7 = in7 + in8; \ + \ + out8 = in7 - in8; \ + out9 = in6 - in9; \ + out10 = in5 - in10; \ + out11 = in4 - in11; \ + out12 = in3 - in12; \ + out13 = in2 - in13; \ + out14 = in1 - in14; \ + out15 = in0 - in15; \ +} + +/* Description : Transposes input 4x4 byte block + Arguments : Inputs - in0, in1, in2, in3 (input 4x4 byte block) + Outputs - out0, out1, out2, out3 (output 4x4 byte block) + Return Type - unsigned byte + Details : +*/ +#define TRANSPOSE4x4_UB_UB(in0, in1, in2, in3, out0, out1, out2, out3) \ +{ \ + v16i8 zero_m = { 0 }; \ + v16i8 s0_m, s1_m, s2_m, s3_m; \ + \ + ILVR_D2_SB(in1, in0, in3, in2, s0_m, s1_m); \ + ILVRL_B2_SB(s1_m, s0_m, s2_m, s3_m); \ + \ + out0 = (v16u8) __msa_ilvr_b(s3_m, s2_m); \ + out1 = (v16u8) __msa_sldi_b(zero_m, (v16i8) out0, 4); \ + out2 = (v16u8) __msa_sldi_b(zero_m, (v16i8) out1, 4); \ + out3 = (v16u8) __msa_sldi_b(zero_m, (v16i8) out2, 4); \ +} + +/* Description : Transposes input 8x4 byte block into 4x8 + Arguments : Inputs - in0, in1, in2, in3 (input 8x4 byte block) + Outputs - out0, out1, out2, out3 (output 4x8 byte block) + Return Type - unsigned byte + Details : +*/ +#define TRANSPOSE8x4_UB(RTYPE, in0, in1, in2, in3, in4, in5, in6, in7, \ + out0, out1, out2, out3) \ +{ \ + v16i8 tmp0_m, tmp1_m, tmp2_m, tmp3_m; \ + \ + ILVEV_W2_SB(in0, in4, in1, in5, tmp0_m, tmp1_m); \ + tmp2_m = __msa_ilvr_b(tmp1_m, tmp0_m); \ + ILVEV_W2_SB(in2, in6, in3, in7, tmp0_m, tmp1_m); \ + \ + tmp3_m = __msa_ilvr_b(tmp1_m, tmp0_m); \ + ILVRL_H2_SB(tmp3_m, tmp2_m, tmp0_m, tmp1_m); \ + \ + ILVRL_W2(RTYPE, tmp1_m, tmp0_m, out0, out2); \ + out1 = (RTYPE) __msa_ilvl_d((v2i64) out2, (v2i64) out0); \ + out3 = (RTYPE) __msa_ilvl_d((v2i64) out0, (v2i64) out2); \ +} +#define TRANSPOSE8x4_UB_UB(...) TRANSPOSE8x4_UB(v16u8, __VA_ARGS__) +#define TRANSPOSE8x4_UB_UH(...) TRANSPOSE8x4_UB(v8u16, __VA_ARGS__) + +/* Description : Transposes input 8x8 byte block + Arguments : Inputs - in0, in1, in2, in3, in4, in5, in6, in7 + (input 8x8 byte block) + Outputs - out0, out1, out2, out3, out4, out5, out6, out7 + (output 8x8 byte block) + Return Type - unsigned byte + Details : +*/ +#define TRANSPOSE8x8_UB(RTYPE, in0, in1, in2, in3, in4, in5, in6, in7, \ + out0, out1, out2, out3, out4, out5, out6, out7) \ +{ \ + v16i8 tmp0_m, tmp1_m, tmp2_m, tmp3_m; \ + v16i8 tmp4_m, tmp5_m, tmp6_m, tmp7_m; \ + \ + ILVR_B4_SB(in2, in0, in3, in1, in6, in4, in7, in5, \ + tmp0_m, tmp1_m, tmp2_m, tmp3_m); \ + ILVRL_B2_SB(tmp1_m, tmp0_m, tmp4_m, tmp5_m); \ + ILVRL_B2_SB(tmp3_m, tmp2_m, tmp6_m, tmp7_m); \ + ILVRL_W2(RTYPE, tmp6_m, tmp4_m, out0, out2); \ + ILVRL_W2(RTYPE, tmp7_m, tmp5_m, out4, out6); \ + SLDI_B2_0(RTYPE, out0, out2, out1, out3, 8); \ + SLDI_B2_0(RTYPE, out4, out6, out5, out7, 8); \ +} +#define TRANSPOSE8x8_UB_UB(...) TRANSPOSE8x8_UB(v16u8, __VA_ARGS__) +#define TRANSPOSE8x8_UB_UH(...) TRANSPOSE8x8_UB(v8u16, __VA_ARGS__) + +/* Description : Transposes 16x4 block into 4x16 with byte elements in vectors + Arguments : Inputs - in0, in1, in2, in3, in4, in5, in6, in7, + in8, in9, in10, in11, in12, in13, in14, in15 + Outputs - out0, out1, out2, out3 + Return Type - unsigned byte + Details : +*/ +#define TRANSPOSE16x4_UB_UB(in0, in1, in2, in3, in4, in5, in6, in7, \ + in8, in9, in10, in11, in12, in13, in14, in15, \ + out0, out1, out2, out3) \ +{ \ + v2i64 tmp0_m, tmp1_m, tmp2_m, tmp3_m; \ + \ + ILVEV_W2_SD(in0, in4, in8, in12, tmp0_m, tmp1_m); \ + out1 = (v16u8) __msa_ilvev_d(tmp1_m, tmp0_m); \ + \ + ILVEV_W2_SD(in1, in5, in9, in13, tmp0_m, tmp1_m); \ + out3 = (v16u8) __msa_ilvev_d(tmp1_m, tmp0_m); \ + \ + ILVEV_W2_SD(in2, in6, in10, in14, tmp0_m, tmp1_m); \ + \ + tmp2_m = __msa_ilvev_d(tmp1_m, tmp0_m); \ + ILVEV_W2_SD(in3, in7, in11, in15, tmp0_m, tmp1_m); \ + \ + tmp3_m = __msa_ilvev_d(tmp1_m, tmp0_m); \ + ILVEV_B2_SD(out1, out3, tmp2_m, tmp3_m, tmp0_m, tmp1_m); \ + out0 = (v16u8) __msa_ilvev_h((v8i16) tmp1_m, (v8i16) tmp0_m); \ + out2 = (v16u8) __msa_ilvod_h((v8i16) tmp1_m, (v8i16) tmp0_m); \ + \ + tmp0_m = (v2i64) __msa_ilvod_b((v16i8) out3, (v16i8) out1); \ + tmp1_m = (v2i64) __msa_ilvod_b((v16i8) tmp3_m, (v16i8) tmp2_m); \ + out1 = (v16u8) __msa_ilvev_h((v8i16) tmp1_m, (v8i16) tmp0_m); \ + out3 = (v16u8) __msa_ilvod_h((v8i16) tmp1_m, (v8i16) tmp0_m); \ +} + +/* Description : Transposes 16x8 block into 8x16 with byte elements in vectors + Arguments : Inputs - in0, in1, in2, in3, in4, in5, in6, in7, + in8, in9, in10, in11, in12, in13, in14, in15 + Outputs - out0, out1, out2, out3, out4, out5, out6, out7 + Return Type - unsigned byte + Details : +*/ +#define TRANSPOSE16x8_UB_UB(in0, in1, in2, in3, in4, in5, in6, in7, \ + in8, in9, in10, in11, in12, in13, in14, in15, \ + out0, out1, out2, out3, out4, out5, out6, out7) \ +{ \ + v16u8 tmp0_m, tmp1_m, tmp2_m, tmp3_m; \ + v16u8 tmp4_m, tmp5_m, tmp6_m, tmp7_m; \ + \ + ILVEV_D2_UB(in0, in8, in1, in9, out7, out6); \ + ILVEV_D2_UB(in2, in10, in3, in11, out5, out4); \ + ILVEV_D2_UB(in4, in12, in5, in13, out3, out2); \ + ILVEV_D2_UB(in6, in14, in7, in15, out1, out0); \ + \ + tmp0_m = (v16u8) __msa_ilvev_b((v16i8) out6, (v16i8) out7); \ + tmp4_m = (v16u8) __msa_ilvod_b((v16i8) out6, (v16i8) out7); \ + tmp1_m = (v16u8) __msa_ilvev_b((v16i8) out4, (v16i8) out5); \ + tmp5_m = (v16u8) __msa_ilvod_b((v16i8) out4, (v16i8) out5); \ + out5 = (v16u8) __msa_ilvev_b((v16i8) out2, (v16i8) out3); \ + tmp6_m = (v16u8) __msa_ilvod_b((v16i8) out2, (v16i8) out3); \ + out7 = (v16u8) __msa_ilvev_b((v16i8) out0, (v16i8) out1); \ + tmp7_m = (v16u8) __msa_ilvod_b((v16i8) out0, (v16i8) out1); \ + \ + ILVEV_H2_UB(tmp0_m, tmp1_m, out5, out7, tmp2_m, tmp3_m); \ + out0 = (v16u8) __msa_ilvev_w((v4i32) tmp3_m, (v4i32) tmp2_m); \ + out4 = (v16u8) __msa_ilvod_w((v4i32) tmp3_m, (v4i32) tmp2_m); \ + \ + tmp2_m = (v16u8) __msa_ilvod_h((v8i16) tmp1_m, (v8i16) tmp0_m); \ + tmp3_m = (v16u8) __msa_ilvod_h((v8i16) out7, (v8i16) out5); \ + out2 = (v16u8) __msa_ilvev_w((v4i32) tmp3_m, (v4i32) tmp2_m); \ + out6 = (v16u8) __msa_ilvod_w((v4i32) tmp3_m, (v4i32) tmp2_m); \ + \ + ILVEV_H2_UB(tmp4_m, tmp5_m, tmp6_m, tmp7_m, tmp2_m, tmp3_m); \ + out1 = (v16u8) __msa_ilvev_w((v4i32) tmp3_m, (v4i32) tmp2_m); \ + out5 = (v16u8) __msa_ilvod_w((v4i32) tmp3_m, (v4i32) tmp2_m); \ + \ + tmp2_m = (v16u8) __msa_ilvod_h((v8i16) tmp5_m, (v8i16) tmp4_m); \ + tmp2_m = (v16u8) __msa_ilvod_h((v8i16) tmp5_m, (v8i16) tmp4_m); \ + tmp3_m = (v16u8) __msa_ilvod_h((v8i16) tmp7_m, (v8i16) tmp6_m); \ + tmp3_m = (v16u8) __msa_ilvod_h((v8i16) tmp7_m, (v8i16) tmp6_m); \ + out3 = (v16u8) __msa_ilvev_w((v4i32) tmp3_m, (v4i32) tmp2_m); \ + out7 = (v16u8) __msa_ilvod_w((v4i32) tmp3_m, (v4i32) tmp2_m); \ +} + +/* Description : Transposes 4x4 block with half word elements in vectors + Arguments : Inputs - in0, in1, in2, in3 + Outputs - out0, out1, out2, out3 + Return Type - signed halfword + Details : +*/ +#define TRANSPOSE4x4_SH_SH(in0, in1, in2, in3, out0, out1, out2, out3) \ +{ \ + v8i16 s0_m, s1_m; \ + \ + ILVR_H2_SH(in1, in0, in3, in2, s0_m, s1_m); \ + ILVRL_W2_SH(s1_m, s0_m, out0, out2); \ + out1 = (v8i16) __msa_ilvl_d((v2i64) out0, (v2i64) out0); \ + out3 = (v8i16) __msa_ilvl_d((v2i64) out0, (v2i64) out2); \ +} + +/* Description : Transposes 8x8 block with half word elements in vectors + Arguments : Inputs - in0, in1, in2, in3, in4, in5, in6, in7 + Outputs - out0, out1, out2, out3, out4, out5, out6, out7 + Return Type - signed halfword + Details : +*/ +#define TRANSPOSE8x8_H(RTYPE, in0, in1, in2, in3, in4, in5, in6, in7, \ + out0, out1, out2, out3, out4, out5, out6, out7) \ +{ \ + v8i16 s0_m, s1_m; \ + v8i16 tmp0_m, tmp1_m, tmp2_m, tmp3_m; \ + v8i16 tmp4_m, tmp5_m, tmp6_m, tmp7_m; \ + \ + ILVR_H2_SH(in6, in4, in7, in5, s0_m, s1_m); \ + ILVRL_H2_SH(s1_m, s0_m, tmp0_m, tmp1_m); \ + ILVL_H2_SH(in6, in4, in7, in5, s0_m, s1_m); \ + ILVRL_H2_SH(s1_m, s0_m, tmp2_m, tmp3_m); \ + ILVR_H2_SH(in2, in0, in3, in1, s0_m, s1_m); \ + ILVRL_H2_SH(s1_m, s0_m, tmp4_m, tmp5_m); \ + ILVL_H2_SH(in2, in0, in3, in1, s0_m, s1_m); \ + ILVRL_H2_SH(s1_m, s0_m, tmp6_m, tmp7_m); \ + PCKEV_D4(RTYPE, tmp0_m, tmp4_m, tmp1_m, tmp5_m, tmp2_m, tmp6_m, \ + tmp3_m, tmp7_m, out0, out2, out4, out6); \ + out1 = (RTYPE) __msa_pckod_d((v2i64) tmp0_m, (v2i64) tmp4_m); \ + out3 = (RTYPE) __msa_pckod_d((v2i64) tmp1_m, (v2i64) tmp5_m); \ + out5 = (RTYPE) __msa_pckod_d((v2i64) tmp2_m, (v2i64) tmp6_m); \ + out7 = (RTYPE) __msa_pckod_d((v2i64) tmp3_m, (v2i64) tmp7_m); \ +} +#define TRANSPOSE8x8_UH_UH(...) TRANSPOSE8x8_H(v8u16, __VA_ARGS__) +#define TRANSPOSE8x8_SH_SH(...) TRANSPOSE8x8_H(v8i16, __VA_ARGS__) + +/* Description : Transposes 4x4 block with word elements in vectors + Arguments : Inputs - in0, in1, in2, in3 + Outputs - out0, out1, out2, out3 + Return Type - signed word + Details : +*/ +#define TRANSPOSE4x4_SW_SW(in0, in1, in2, in3, out0, out1, out2, out3) \ +{ \ + v4i32 s0_m, s1_m, s2_m, s3_m; \ + \ + ILVRL_W2_SW(in1, in0, s0_m, s1_m); \ + ILVRL_W2_SW(in3, in2, s2_m, s3_m); \ + \ + out0 = (v4i32) __msa_ilvr_d((v2i64) s2_m, (v2i64) s0_m); \ + out1 = (v4i32) __msa_ilvl_d((v2i64) s2_m, (v2i64) s0_m); \ + out2 = (v4i32) __msa_ilvr_d((v2i64) s3_m, (v2i64) s1_m); \ + out3 = (v4i32) __msa_ilvl_d((v2i64) s3_m, (v2i64) s1_m); \ +} + +/* Description : Average byte elements from pair of vectors and store 8x4 byte + block in destination memory + Arguments : Inputs - in0, in1, in2, in3, in4, in5, in6, in7, pdst, stride + Outputs - + Return Type - + Details : Each byte element from input vector pair 'in0' and 'in1' are + averaged (a + b)/2 and stored in 'tmp0_m' + Each byte element from input vector pair 'in2' and 'in3' are + averaged (a + b)/2 and stored in 'tmp1_m' + Each byte element from input vector pair 'in4' and 'in5' are + averaged (a + b)/2 and stored in 'tmp2_m' + Each byte element from input vector pair 'in6' and 'in7' are + averaged (a + b)/2 and stored in 'tmp3_m' + The half vector results from all 4 vectors are stored in + destination memory as 8x4 byte block +*/ +#define AVE_ST8x4_UB(in0, in1, in2, in3, in4, in5, in6, in7, pdst, stride) \ +{ \ + uint64_t out0_m, out1_m, out2_m, out3_m; \ + v16u8 tmp0_m, tmp1_m, tmp2_m, tmp3_m; \ + \ + tmp0_m = __msa_ave_u_b((v16u8) in0, (v16u8) in1); \ + tmp1_m = __msa_ave_u_b((v16u8) in2, (v16u8) in3); \ + tmp2_m = __msa_ave_u_b((v16u8) in4, (v16u8) in5); \ + tmp3_m = __msa_ave_u_b((v16u8) in6, (v16u8) in7); \ + \ + out0_m = __msa_copy_u_d((v2i64) tmp0_m, 0); \ + out1_m = __msa_copy_u_d((v2i64) tmp1_m, 0); \ + out2_m = __msa_copy_u_d((v2i64) tmp2_m, 0); \ + out3_m = __msa_copy_u_d((v2i64) tmp3_m, 0); \ + SD4(out0_m, out1_m, out2_m, out3_m, pdst, stride); \ +} + +/* Description : Average byte elements from pair of vectors and store 16x4 byte + block in destination memory + Arguments : Inputs - in0, in1, in2, in3, in4, in5, in6, in7, pdst, stride + Outputs - + Return Type - + Details : Each byte element from input vector pair 'in0' and 'in1' are + averaged (a + b)/2 and stored in 'tmp0_m' + Each byte element from input vector pair 'in2' and 'in3' are + averaged (a + b)/2 and stored in 'tmp1_m' + Each byte element from input vector pair 'in4' and 'in5' are + averaged (a + b)/2 and stored in 'tmp2_m' + Each byte element from input vector pair 'in6' and 'in7' are + averaged (a + b)/2 and stored in 'tmp3_m' + The results from all 4 vectors are stored in destination + memory as 16x4 byte block +*/ +#define AVE_ST16x4_UB(in0, in1, in2, in3, in4, in5, in6, in7, pdst, stride) \ +{ \ + v16u8 tmp0_m, tmp1_m, tmp2_m, tmp3_m; \ + \ + tmp0_m = __msa_ave_u_b((v16u8) in0, (v16u8) in1); \ + tmp1_m = __msa_ave_u_b((v16u8) in2, (v16u8) in3); \ + tmp2_m = __msa_ave_u_b((v16u8) in4, (v16u8) in5); \ + tmp3_m = __msa_ave_u_b((v16u8) in6, (v16u8) in7); \ + \ + ST_UB4(tmp0_m, tmp1_m, tmp2_m, tmp3_m, pdst, stride); \ +} + +/* Description : Average rounded byte elements from pair of vectors and store + 8x4 byte block in destination memory + Arguments : Inputs - in0, in1, in2, in3, in4, in5, in6, in7, pdst, stride + Outputs - + Return Type - + Details : Each byte element from input vector pair 'in0' and 'in1' are + average rounded (a + b + 1)/2 and stored in 'tmp0_m' + Each byte element from input vector pair 'in2' and 'in3' are + average rounded (a + b + 1)/2 and stored in 'tmp1_m' + Each byte element from input vector pair 'in4' and 'in5' are + average rounded (a + b + 1)/2 and stored in 'tmp2_m' + Each byte element from input vector pair 'in6' and 'in7' are + average rounded (a + b + 1)/2 and stored in 'tmp3_m' + The half vector results from all 4 vectors are stored in + destination memory as 8x4 byte block +*/ +#define AVER_ST8x4_UB(in0, in1, in2, in3, in4, in5, in6, in7, pdst, stride) \ +{ \ + uint64_t out0_m, out1_m, out2_m, out3_m; \ + v16u8 tp0_m, tp1_m, tp2_m, tp3_m; \ + \ + AVER_UB4_UB(in0, in1, in2, in3, in4, in5, in6, in7, \ + tp0_m, tp1_m, tp2_m, tp3_m); \ + \ + out0_m = __msa_copy_u_d((v2i64) tp0_m, 0); \ + out1_m = __msa_copy_u_d((v2i64) tp1_m, 0); \ + out2_m = __msa_copy_u_d((v2i64) tp2_m, 0); \ + out3_m = __msa_copy_u_d((v2i64) tp3_m, 0); \ + SD4(out0_m, out1_m, out2_m, out3_m, pdst, stride); \ +} + +/* Description : Average rounded byte elements from pair of vectors and store + 16x4 byte block in destination memory + Arguments : Inputs - in0, in1, in2, in3, in4, in5, in6, in7, pdst, stride + Outputs - + Return Type - + Details : Each byte element from input vector pair 'in0' and 'in1' are + average rounded (a + b + 1)/2 and stored in 'tmp0_m' + Each byte element from input vector pair 'in2' and 'in3' are + average rounded (a + b + 1)/2 and stored in 'tmp1_m' + Each byte element from input vector pair 'in4' and 'in5' are + average rounded (a + b + 1)/2 and stored in 'tmp2_m' + Each byte element from input vector pair 'in6' and 'in7' are + average rounded (a + b + 1)/2 and stored in 'tmp3_m' + The vector results from all 4 vectors are stored in + destination memory as 16x4 byte block +*/ +#define AVER_ST16x4_UB(in0, in1, in2, in3, in4, in5, in6, in7, pdst, stride) \ +{ \ + v16u8 t0_m, t1_m, t2_m, t3_m; \ + \ + AVER_UB4_UB(in0, in1, in2, in3, in4, in5, in6, in7, \ + t0_m, t1_m, t2_m, t3_m); \ + ST_UB4(t0_m, t1_m, t2_m, t3_m, pdst, stride); \ +} + +/* Description : Average rounded byte elements from pair of vectors, + average rounded with destination and store 8x4 byte block + in destination memory + Arguments : Inputs - in0, in1, in2, in3, in4, in5, in6, in7, pdst, stride + Outputs - + Return Type - + Details : Each byte element from input vector pair 'in0' and 'in1' are + average rounded (a + b + 1)/2 and stored in 'tmp0_m' + Each byte element from input vector pair 'in2' and 'in3' are + average rounded (a + b + 1)/2 and stored in 'tmp1_m' + Each byte element from input vector pair 'in4' and 'in5' are + average rounded (a + b + 1)/2 and stored in 'tmp2_m' + Each byte element from input vector pair 'in6' and 'in7' are + average rounded (a + b + 1)/2 and stored in 'tmp3_m' + The half vector results from all 4 vectors are stored in + destination memory as 8x4 byte block +*/ +#define AVER_DST_ST8x4_UB(in0, in1, in2, in3, in4, in5, in6, in7, \ + pdst, stride) \ +{ \ + v16u8 tmp0_m, tmp1_m, tmp2_m, tmp3_m; \ + v16u8 dst0_m, dst1_m, dst2_m, dst3_m; \ + \ + LD_UB4(pdst, stride, dst0_m, dst1_m, dst2_m, dst3_m); \ + AVER_UB4_UB(in0, in1, in2, in3, in4, in5, in6, in7, \ + tmp0_m, tmp1_m, tmp2_m, tmp3_m); \ + AVER_ST8x4_UB(dst0_m, tmp0_m, dst1_m, tmp1_m, \ + dst2_m, tmp2_m, dst3_m, tmp3_m, pdst, stride); \ +} + +/* Description : Average rounded byte elements from pair of vectors, + average rounded with destination and store 16x4 byte block + in destination memory + Arguments : Inputs - in0, in1, in2, in3, in4, in5, in6, in7, pdst, stride + Outputs - + Return Type - + Details : Each byte element from input vector pair 'in0' and 'in1' are + average rounded (a + b + 1)/2 and stored in 'tmp0_m' + Each byte element from input vector pair 'in2' and 'in3' are + average rounded (a + b + 1)/2 and stored in 'tmp1_m' + Each byte element from input vector pair 'in4' and 'in5' are + average rounded (a + b + 1)/2 and stored in 'tmp2_m' + Each byte element from input vector pair 'in6' and 'in7' are + average rounded (a + b + 1)/2 and stored in 'tmp3_m' + The vector results from all 4 vectors are stored in + destination memory as 16x4 byte block +*/ +#define AVER_DST_ST16x4_UB(in0, in1, in2, in3, in4, in5, in6, in7, \ + pdst, stride) \ +{ \ + v16u8 tmp0_m, tmp1_m, tmp2_m, tmp3_m; \ + v16u8 dst0_m, dst1_m, dst2_m, dst3_m; \ + \ + LD_UB4(pdst, stride, dst0_m, dst1_m, dst2_m, dst3_m); \ + AVER_UB4_UB(in0, in1, in2, in3, in4, in5, in6, in7, \ + tmp0_m, tmp1_m, tmp2_m, tmp3_m); \ + AVER_ST16x4_UB(dst0_m, tmp0_m, dst1_m, tmp1_m, \ + dst2_m, tmp2_m, dst3_m, tmp3_m, pdst, stride); \ +} + +/* Description : Add block 4x4 + Arguments : Inputs - in0, in1, in2, in3, pdst, stride + Outputs - + Return Type - unsigned bytes + Details : Least significant 4 bytes from each input vector are added to + the destination bytes, clipped between 0-255 and then stored. +*/ +#define ADDBLK_ST4x4_UB(in0, in1, in2, in3, pdst, stride) \ +{ \ + uint32_t src0_m, src1_m, src2_m, src3_m; \ + uint32_t out0_m, out1_m, out2_m, out3_m; \ + v8i16 inp0_m, inp1_m, res0_m, res1_m; \ + v16i8 dst0_m = { 0 }; \ + v16i8 dst1_m = { 0 }; \ + v16i8 zero_m = { 0 }; \ + \ + ILVR_D2_SH(in1, in0, in3, in2, inp0_m, inp1_m) \ + LW4(pdst, stride, src0_m, src1_m, src2_m, src3_m); \ + INSERT_W2_SB(src0_m, src1_m, dst0_m); \ + INSERT_W2_SB(src2_m, src3_m, dst1_m); \ + ILVR_B2_SH(zero_m, dst0_m, zero_m, dst1_m, res0_m, res1_m); \ + ADD2(res0_m, inp0_m, res1_m, inp1_m, res0_m, res1_m); \ + CLIP_SH2_0_255(res0_m, res1_m); \ + PCKEV_B2_SB(res0_m, res0_m, res1_m, res1_m, dst0_m, dst1_m); \ + \ + out0_m = __msa_copy_u_w((v4i32) dst0_m, 0); \ + out1_m = __msa_copy_u_w((v4i32) dst0_m, 1); \ + out2_m = __msa_copy_u_w((v4i32) dst1_m, 0); \ + out3_m = __msa_copy_u_w((v4i32) dst1_m, 1); \ + SW4(out0_m, out1_m, out2_m, out3_m, pdst, stride); \ +} + +/* Description : Dot product and addition of 3 signed halfword input vectors + Arguments : Inputs - in0, in1, in2, coeff0, coeff1, coeff2 + Outputs - out0_m + Return Type - signed halfword + Details : Dot product of 'in0' with 'coeff0' + Dot product of 'in1' with 'coeff1' + Dot product of 'in2' with 'coeff2' + Addition of all the 3 vector results + + out0_m = (in0 * coeff0) + (in1 * coeff1) + (in2 * coeff2) +*/ +#define DPADD_SH3_SH(in0, in1, in2, coeff0, coeff1, coeff2) \ +( { \ + v8i16 tmp1_m; \ + v8i16 out0_m; \ + \ + out0_m = __msa_dotp_s_h((v16i8) in0, (v16i8) coeff0); \ + out0_m = __msa_dpadd_s_h(out0_m, (v16i8) in1, (v16i8) coeff1); \ + tmp1_m = __msa_dotp_s_h((v16i8) in2, (v16i8) coeff2); \ + out0_m = __msa_adds_s_h(out0_m, tmp1_m); \ + \ + out0_m; \ +} ) + +/* Description : Pack even elements of input vectors & xor with 128 + Arguments : Inputs - in0, in1 + Outputs - out_m + Return Type - unsigned byte + Details : Signed byte even elements from 'in0' and 'in1' are packed + together in one vector and the resulted vector is xor'ed with + 128 to shift the range from signed to unsigned byte +*/ +#define PCKEV_XORI128_UB(in0, in1) \ +( { \ + v16u8 out_m; \ + out_m = (v16u8) __msa_pckev_b((v16i8) in1, (v16i8) in0); \ + out_m = (v16u8) __msa_xori_b((v16u8) out_m, 128); \ + out_m; \ +} ) + +/* Description : Converts inputs to unsigned bytes, interleave, average & store + as 8x4 unsigned byte block + Arguments : Inputs - in0, in1, in2, in3, dst0, dst1, dst2, dst3, + pdst, stride +*/ +#define CONVERT_UB_AVG_ST8x4_UB(in0, in1, in2, in3, \ + dst0, dst1, dst2, dst3, pdst, stride) \ +{ \ + v16u8 tmp0_m, tmp1_m, tmp2_m, tmp3_m; \ + uint8_t *pdst_m = (uint8_t *) (pdst); \ + \ + tmp0_m = PCKEV_XORI128_UB(in0, in1); \ + tmp1_m = PCKEV_XORI128_UB(in2, in3); \ + ILVR_D2_UB(dst1, dst0, dst3, dst2, tmp2_m, tmp3_m); \ + AVER_UB2_UB(tmp0_m, tmp2_m, tmp1_m, tmp3_m, tmp0_m, tmp1_m); \ + ST8x4_UB(tmp0_m, tmp1_m, pdst_m, stride); \ +} + +/* Description : Pack even byte elements, extract 0 & 2 index words from pair + of results and store 4 words in destination memory as per + stride + Arguments : Inputs - in0, in1, in2, in3, pdst, stride +*/ +#define PCKEV_ST4x4_UB(in0, in1, in2, in3, pdst, stride) \ +{ \ + uint32_t out0_m, out1_m, out2_m, out3_m; \ + v16i8 tmp0_m, tmp1_m; \ + \ + PCKEV_B2_SB(in1, in0, in3, in2, tmp0_m, tmp1_m); \ + \ + out0_m = __msa_copy_u_w((v4i32) tmp0_m, 0); \ + out1_m = __msa_copy_u_w((v4i32) tmp0_m, 2); \ + out2_m = __msa_copy_u_w((v4i32) tmp1_m, 0); \ + out3_m = __msa_copy_u_w((v4i32) tmp1_m, 2); \ + \ + SW4(out0_m, out1_m, out2_m, out3_m, pdst, stride); \ +} + +/* Description : Pack even byte elements and store byte vector in destination + memory + Arguments : Inputs - in0, in1, pdst +*/ +#define PCKEV_ST_SB(in0, in1, pdst) \ +{ \ + v16i8 tmp_m; \ + tmp_m = __msa_pckev_b((v16i8) in1, (v16i8) in0); \ + ST_SB(tmp_m, (pdst)); \ +} + +/* Description : Horizontal 2 tap filter kernel code + Arguments : Inputs - in0, in1, mask, coeff, shift +*/ +#define HORIZ_2TAP_FILT_UH(in0, in1, mask, coeff, shift) \ +( { \ + v16i8 tmp0_m; \ + v8u16 tmp1_m; \ + \ + tmp0_m = __msa_vshf_b((v16i8) mask, (v16i8) in1, (v16i8) in0); \ + tmp1_m = __msa_dotp_u_h((v16u8) tmp0_m, (v16u8) coeff); \ + tmp1_m = (v8u16) __msa_srari_h((v8i16) tmp1_m, shift); \ + tmp1_m = __msa_sat_u_h(tmp1_m, shift); \ + \ + tmp1_m; \ +} ) +#endif /* AVUTIL_MIPS_GENERIC_MACROS_MSA_H */ diff --git a/libavutil/mips/intreadwrite.h b/libavutil/mips/intreadwrite.h index 0db3da5666..32084f68f6 100644 --- a/libavutil/mips/intreadwrite.h +++ b/libavutil/mips/intreadwrite.h @@ -1,20 +1,20 @@ /* * Copyright (c) 2009 Mans Rullgard <mans@mansr.com> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -24,13 +24,7 @@ #include <stdint.h> #include "config.h" -/* - * GCC actually handles unaligned accesses correctly in all cases - * except, absurdly, 32-bit loads on mips64. - * - * https://git.libav.org/?p=libav.git;a=commit;h=b82b49a5b774b6ad9119e981c72b8f594fee2ae0 - */ -#if HAVE_MIPS64R2_INLINE || HAVE_MIPS64R1_INLINE +#if ARCH_MIPS64 && HAVE_INLINE_ASM && !HAVE_MIPS64R6 #define AV_RN32 AV_RN32 static av_always_inline uint32_t AV_RN32(const void *p) @@ -47,6 +41,6 @@ static av_always_inline uint32_t AV_RN32(const void *p) return v; } -#endif /* HAVE_MIPS64R2_INLINE || HAVE_MIPS64R1_INLINE */ +#endif /* ARCH_MIPS64 && HAVE_INLINE_ASM && !HAVE_MIPS64R6 */ #endif /* AVUTIL_MIPS_INTREADWRITE_H */ diff --git a/libavutil/mips/libm_mips.h b/libavutil/mips/libm_mips.h new file mode 100644 index 0000000000..757867bce9 --- /dev/null +++ b/libavutil/mips/libm_mips.h @@ -0,0 +1,73 @@ +/* + * Copyright (c) 2012 + * MIPS Technologies, Inc., California. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * 3. Neither the name of the MIPS Technologies, Inc., nor the names of its + * contributors may be used to endorse or promote products derived from + * this software without specific prior written permission. + * + * THIS SOFTWARE IS PROVIDED BY THE MIPS TECHNOLOGIES, INC. ``AS IS'' AND + * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE + * ARE DISCLAIMED. IN NO EVENT SHALL THE MIPS TECHNOLOGIES, INC. BE LIABLE + * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT + * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY + * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF + * SUCH DAMAGE. + * + * Author: Nedeljko Babic (nbabic@mips.com) + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +/** + * @file + * MIPS optimization for some libm functions + */ + +#ifndef AVUTIL_MIPS_LIBM_MIPS_H +#define AVUTIL_MIPS_LIBM_MIPS_H + +static av_always_inline av_const long int lrintf_mips(float x) +{ + register int ret_int; + + __asm__ volatile ( + "cvt.w.s %[x], %[x] \n\t" + "mfc1 %[ret_int], %[x] \n\t" + + :[x]"+f"(x), [ret_int]"=r"(ret_int) + ); + return ret_int; +} + +#undef lrintf +#define lrintf(x) lrintf_mips(x) + +#define HAVE_LRINTF 1 +#endif /* AVUTIL_MIPS_LIBM_MIPS_H */ diff --git a/libavutil/motion_vector.h b/libavutil/motion_vector.h new file mode 100644 index 0000000000..ec29556388 --- /dev/null +++ b/libavutil/motion_vector.h @@ -0,0 +1,57 @@ +/* + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#ifndef AVUTIL_MOTION_VECTOR_H +#define AVUTIL_MOTION_VECTOR_H + +#include <stdint.h> + +typedef struct AVMotionVector { + /** + * Where the current macroblock comes from; negative value when it comes + * from the past, positive value when it comes from the future. + * XXX: set exact relative ref frame reference instead of a +/- 1 "direction". + */ + int32_t source; + /** + * Width and height of the block. + */ + uint8_t w, h; + /** + * Absolute source position. Can be outside the frame area. + */ + int16_t src_x, src_y; + /** + * Absolute destination position. Can be outside the frame area. + */ + int16_t dst_x, dst_y; + /** + * Extra flag information. + * Currently unused. + */ + uint64_t flags; + /** + * Motion vector + * src_x = dst_x + motion_x / motion_scale + * src_y = dst_y + motion_y / motion_scale + */ + int32_t motion_x, motion_y; + uint16_t motion_scale; +} AVMotionVector; + +#endif /* AVUTIL_MOTION_VECTOR_H */ diff --git a/libavutil/murmur3-test.c b/libavutil/murmur3-test.c new file mode 100644 index 0000000000..19ea0a46ff --- /dev/null +++ b/libavutil/murmur3-test.c @@ -0,0 +1,56 @@ +/* + * Copyright (C) 2013 Reimar Döffinger <Reimar.Doeffinger@gmx.de> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "murmur3.c" + +int main(void) +{ + int i; + uint8_t hash_result[16] = {0}; + AVMurMur3 *ctx = av_murmur3_alloc(); +#if 1 + uint8_t in[256] = {0}; + uint8_t *hashes = av_mallocz(256 * 16); + for (i = 0; i < 256; i++) + { + in[i] = i; + av_murmur3_init_seeded(ctx, 256 - i); + // Note: this actually tests hashing 0 bytes + av_murmur3_update(ctx, in, i); + av_murmur3_final(ctx, hashes + 16 * i); + } + av_murmur3_init_seeded(ctx, 0); + av_murmur3_update(ctx, hashes, 256 * 16); + av_murmur3_final(ctx, hash_result); + av_free(hashes); + av_freep(&ctx); + printf("result: 0x%"PRIx64" 0x%"PRIx64"\n", AV_RL64(hash_result), AV_RL64(hash_result + 8)); + // official reference value is 32 bit + return AV_RL32(hash_result) != 0x6384ba69; +#else + uint8_t *in = av_mallocz(512*1024); + av_murmur3_init(ctx); + for (i = 0; i < 40*1024; i++) + av_murmur3_update(ctx, in, 512*1024); + av_murmur3_final(ctx, hash_result); + av_free(in); + return hash_result[0]; +#endif +} diff --git a/libavutil/murmur3.c b/libavutil/murmur3.c new file mode 100644 index 0000000000..4271e01453 --- /dev/null +++ b/libavutil/murmur3.c @@ -0,0 +1,155 @@ +/* + * Copyright (C) 2013 Reimar Döffinger <Reimar.Doeffinger@gmx.de> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ +#include <stdint.h> +#include "mem.h" +#include "intreadwrite.h" +#include "murmur3.h" + +typedef struct AVMurMur3 { + uint64_t h1, h2; + uint8_t state[16]; + int state_pos; + uint64_t len; +} AVMurMur3; + +AVMurMur3 *av_murmur3_alloc(void) +{ + return av_mallocz(sizeof(AVMurMur3)); +} + +void av_murmur3_init_seeded(AVMurMur3 *c, uint64_t seed) +{ + memset(c, 0, sizeof(*c)); + c->h1 = c->h2 = seed; +} + +void av_murmur3_init(AVMurMur3 *c) +{ + // arbitrary random number as seed + av_murmur3_init_seeded(c, 0x725acc55daddca55); +} + +static const uint64_t c1 = UINT64_C(0x87c37b91114253d5); +static const uint64_t c2 = UINT64_C(0x4cf5ad432745937f); + +#define ROT(a, b) (((a) << (b)) | ((a) >> (64 - (b)))) + +static uint64_t inline get_k1(const uint8_t *src) +{ + uint64_t k = AV_RL64(src); + k *= c1; + k = ROT(k, 31); + k *= c2; + return k; +} + +static uint64_t inline get_k2(const uint8_t *src) +{ + uint64_t k = AV_RL64(src + 8); + k *= c2; + k = ROT(k, 33); + k *= c1; + return k; +} + +static uint64_t inline update_h1(uint64_t k, uint64_t h1, uint64_t h2) +{ + k ^= h1; + k = ROT(k, 27); + k += h2; + k *= 5; + k += 0x52dce729; + return k; +} + +static uint64_t inline update_h2(uint64_t k, uint64_t h1, uint64_t h2) +{ + k ^= h2; + k = ROT(k, 31); + k += h1; + k *= 5; + k += 0x38495ab5; + return k; +} + +void av_murmur3_update(AVMurMur3 *c, const uint8_t *src, int len) +{ + const uint8_t *end; + uint64_t h1 = c->h1, h2 = c->h2; + uint64_t k1, k2; + if (len <= 0) return; + c->len += len; + if (c->state_pos > 0) { + while (c->state_pos < 16) { + c->state[c->state_pos++] = *src++; + if (--len <= 0) return; + } + c->state_pos = 0; + k1 = get_k1(c->state); + k2 = get_k2(c->state); + h1 = update_h1(k1, h1, h2); + h2 = update_h2(k2, h1, h2); + } + + end = src + (len & ~15); + while (src < end) { + // These could be done sequentially instead + // of interleaved, but like this is over 10% faster + k1 = get_k1(src); + k2 = get_k2(src); + h1 = update_h1(k1, h1, h2); + h2 = update_h2(k2, h1, h2); + src += 16; + } + c->h1 = h1; + c->h2 = h2; + + len &= 15; + if (len > 0) { + memcpy(c->state, src, len); + c->state_pos = len; + } +} + +static inline uint64_t fmix(uint64_t k) +{ + k ^= k >> 33; + k *= UINT64_C(0xff51afd7ed558ccd); + k ^= k >> 33; + k *= UINT64_C(0xc4ceb9fe1a85ec53); + k ^= k >> 33; + return k; +} + +void av_murmur3_final(AVMurMur3 *c, uint8_t dst[16]) +{ + uint64_t h1 = c->h1, h2 = c->h2; + memset(c->state + c->state_pos, 0, sizeof(c->state) - c->state_pos); + h1 ^= get_k1(c->state) ^ c->len; + h2 ^= get_k2(c->state) ^ c->len; + h1 += h2; + h2 += h1; + h1 = fmix(h1); + h2 = fmix(h2); + h1 += h2; + h2 += h1; + AV_WL64(dst, h1); + AV_WL64(dst + 8, h2); +} diff --git a/libavutil/murmur3.h b/libavutil/murmur3.h new file mode 100644 index 0000000000..f29ed973e9 --- /dev/null +++ b/libavutil/murmur3.h @@ -0,0 +1,32 @@ +/* + * Copyright (C) 2013 Reimar Döffinger <Reimar.Doeffinger@gmx.de> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#ifndef AVUTIL_MURMUR3_H +#define AVUTIL_MURMUR3_H + +#include <stdint.h> + +struct AVMurMur3 *av_murmur3_alloc(void); +void av_murmur3_init_seeded(struct AVMurMur3 *c, uint64_t seed); +void av_murmur3_init(struct AVMurMur3 *c); +void av_murmur3_update(struct AVMurMur3 *c, const uint8_t *src, int len); +void av_murmur3_final(struct AVMurMur3 *c, uint8_t dst[16]); + +#endif /* AVUTIL_MURMUR3_H */ diff --git a/libavutil/opencl.c b/libavutil/opencl.c new file mode 100644 index 0000000000..08b50341c0 --- /dev/null +++ b/libavutil/opencl.c @@ -0,0 +1,875 @@ +/* + * Copyright (C) 2012 Peng Gao <peng@multicorewareinc.com> + * Copyright (C) 2012 Li Cao <li@multicorewareinc.com> + * Copyright (C) 2012 Wei Gao <weigao@multicorewareinc.com> + * Copyright (C) 2013 Lenny Wang <lwanghpc@gmail.com> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "opencl.h" +#include "avstring.h" +#include "log.h" +#include "avassert.h" +#include "opt.h" + +#if HAVE_THREADS +#include "thread.h" +#include "atomic.h" + +static pthread_mutex_t * volatile atomic_opencl_lock = NULL; +#define LOCK_OPENCL pthread_mutex_lock(atomic_opencl_lock) +#define UNLOCK_OPENCL pthread_mutex_unlock(atomic_opencl_lock) +#else +#define LOCK_OPENCL +#define UNLOCK_OPENCL +#endif + +#define MAX_KERNEL_CODE_NUM 200 + +typedef struct { + int is_compiled; + const char *kernel_string; +} KernelCode; + +typedef struct { + const AVClass *class; + int log_offset; + void *log_ctx; + int init_count; + int opt_init_flag; + /** + * if set to 1, the OpenCL environment was created by the user and + * passed as AVOpenCLExternalEnv when initing ,0:created by opencl wrapper. + */ + int is_user_created; + int platform_idx; + int device_idx; + cl_platform_id platform_id; + cl_device_type device_type; + cl_context context; + cl_device_id device_id; + cl_command_queue command_queue; + int kernel_code_count; + KernelCode kernel_code[MAX_KERNEL_CODE_NUM]; + AVOpenCLDeviceList device_list; +} OpenclContext; + +#define OFFSET(x) offsetof(OpenclContext, x) + +static const AVOption opencl_options[] = { + { "platform_idx", "set platform index value", OFFSET(platform_idx), AV_OPT_TYPE_INT, {.i64=-1}, -1, INT_MAX}, + { "device_idx", "set device index value", OFFSET(device_idx), AV_OPT_TYPE_INT, {.i64=-1}, -1, INT_MAX}, + { NULL } +}; + +static const AVClass openclutils_class = { + .class_name = "opencl", + .option = opencl_options, + .item_name = av_default_item_name, + .version = LIBAVUTIL_VERSION_INT, + .log_level_offset_offset = offsetof(OpenclContext, log_offset), + .parent_log_context_offset = offsetof(OpenclContext, log_ctx), +}; + +static OpenclContext opencl_ctx = {&openclutils_class}; + +static const cl_device_type device_type[] = {CL_DEVICE_TYPE_GPU, CL_DEVICE_TYPE_CPU}; + +typedef struct { + int err_code; + const char *err_str; +} OpenclErrorMsg; + +static const OpenclErrorMsg opencl_err_msg[] = { + {CL_DEVICE_NOT_FOUND, "DEVICE NOT FOUND"}, + {CL_DEVICE_NOT_AVAILABLE, "DEVICE NOT AVAILABLE"}, + {CL_COMPILER_NOT_AVAILABLE, "COMPILER NOT AVAILABLE"}, + {CL_MEM_OBJECT_ALLOCATION_FAILURE, "MEM OBJECT ALLOCATION FAILURE"}, + {CL_OUT_OF_RESOURCES, "OUT OF RESOURCES"}, + {CL_OUT_OF_HOST_MEMORY, "OUT OF HOST MEMORY"}, + {CL_PROFILING_INFO_NOT_AVAILABLE, "PROFILING INFO NOT AVAILABLE"}, + {CL_MEM_COPY_OVERLAP, "MEM COPY OVERLAP"}, + {CL_IMAGE_FORMAT_MISMATCH, "IMAGE FORMAT MISMATCH"}, + {CL_IMAGE_FORMAT_NOT_SUPPORTED, "IMAGE FORMAT NOT_SUPPORTED"}, + {CL_BUILD_PROGRAM_FAILURE, "BUILD PROGRAM FAILURE"}, + {CL_MAP_FAILURE, "MAP FAILURE"}, + {CL_MISALIGNED_SUB_BUFFER_OFFSET, "MISALIGNED SUB BUFFER OFFSET"}, + {CL_EXEC_STATUS_ERROR_FOR_EVENTS_IN_WAIT_LIST, "EXEC STATUS ERROR FOR EVENTS IN WAIT LIST"}, + {CL_COMPILE_PROGRAM_FAILURE, "COMPILE PROGRAM FAILURE"}, + {CL_LINKER_NOT_AVAILABLE, "LINKER NOT AVAILABLE"}, + {CL_LINK_PROGRAM_FAILURE, "LINK PROGRAM FAILURE"}, + {CL_DEVICE_PARTITION_FAILED, "DEVICE PARTITION FAILED"}, + {CL_KERNEL_ARG_INFO_NOT_AVAILABLE, "KERNEL ARG INFO NOT AVAILABLE"}, + {CL_INVALID_VALUE, "INVALID VALUE"}, + {CL_INVALID_DEVICE_TYPE, "INVALID DEVICE TYPE"}, + {CL_INVALID_PLATFORM, "INVALID PLATFORM"}, + {CL_INVALID_DEVICE, "INVALID DEVICE"}, + {CL_INVALID_CONTEXT, "INVALID CONTEXT"}, + {CL_INVALID_QUEUE_PROPERTIES, "INVALID QUEUE PROPERTIES"}, + {CL_INVALID_COMMAND_QUEUE, "INVALID COMMAND QUEUE"}, + {CL_INVALID_HOST_PTR, "INVALID HOST PTR"}, + {CL_INVALID_MEM_OBJECT, "INVALID MEM OBJECT"}, + {CL_INVALID_IMAGE_FORMAT_DESCRIPTOR, "INVALID IMAGE FORMAT DESCRIPTOR"}, + {CL_INVALID_IMAGE_SIZE, "INVALID IMAGE SIZE"}, + {CL_INVALID_SAMPLER, "INVALID SAMPLER"}, + {CL_INVALID_BINARY, "INVALID BINARY"}, + {CL_INVALID_BUILD_OPTIONS, "INVALID BUILD OPTIONS"}, + {CL_INVALID_PROGRAM, "INVALID PROGRAM"}, + {CL_INVALID_PROGRAM_EXECUTABLE, "INVALID PROGRAM EXECUTABLE"}, + {CL_INVALID_KERNEL_NAME, "INVALID KERNEL NAME"}, + {CL_INVALID_KERNEL_DEFINITION, "INVALID KERNEL DEFINITION"}, + {CL_INVALID_KERNEL, "INVALID KERNEL"}, + {CL_INVALID_ARG_INDEX, "INVALID ARG INDEX"}, + {CL_INVALID_ARG_VALUE, "INVALID ARG VALUE"}, + {CL_INVALID_ARG_SIZE, "INVALID ARG_SIZE"}, + {CL_INVALID_KERNEL_ARGS, "INVALID KERNEL ARGS"}, + {CL_INVALID_WORK_DIMENSION, "INVALID WORK DIMENSION"}, + {CL_INVALID_WORK_GROUP_SIZE, "INVALID WORK GROUP SIZE"}, + {CL_INVALID_WORK_ITEM_SIZE, "INVALID WORK ITEM SIZE"}, + {CL_INVALID_GLOBAL_OFFSET, "INVALID GLOBAL OFFSET"}, + {CL_INVALID_EVENT_WAIT_LIST, "INVALID EVENT WAIT LIST"}, + {CL_INVALID_EVENT, "INVALID EVENT"}, + {CL_INVALID_OPERATION, "INVALID OPERATION"}, + {CL_INVALID_GL_OBJECT, "INVALID GL OBJECT"}, + {CL_INVALID_BUFFER_SIZE, "INVALID BUFFER SIZE"}, + {CL_INVALID_MIP_LEVEL, "INVALID MIP LEVEL"}, + {CL_INVALID_GLOBAL_WORK_SIZE, "INVALID GLOBAL WORK SIZE"}, + {CL_INVALID_PROPERTY, "INVALID PROPERTY"}, + {CL_INVALID_IMAGE_DESCRIPTOR, "INVALID IMAGE DESCRIPTOR"}, + {CL_INVALID_COMPILER_OPTIONS, "INVALID COMPILER OPTIONS"}, + {CL_INVALID_LINKER_OPTIONS, "INVALID LINKER OPTIONS"}, + {CL_INVALID_DEVICE_PARTITION_COUNT, "INVALID DEVICE PARTITION COUNT"}, +}; + +const char *av_opencl_errstr(cl_int status) +{ + int i; + for (i = 0; i < FF_ARRAY_ELEMS(opencl_err_msg); i++) { + if (opencl_err_msg[i].err_code == status) + return opencl_err_msg[i].err_str; + } + return "unknown error"; +} + +static void free_device_list(AVOpenCLDeviceList *device_list) +{ + int i, j; + if (!device_list) + return; + for (i = 0; i < device_list->platform_num; i++) { + if (!device_list->platform_node[i]) + continue; + for (j = 0; j < device_list->platform_node[i]->device_num; j++) { + av_freep(&(device_list->platform_node[i]->device_node[j]->device_name)); + av_freep(&(device_list->platform_node[i]->device_node[j])); + } + av_freep(&device_list->platform_node[i]->device_node); + av_freep(&(device_list->platform_node[i]->platform_name)); + av_freep(&device_list->platform_node[i]); + } + av_freep(&device_list->platform_node); + device_list->platform_num = 0; +} + +static int get_device_list(AVOpenCLDeviceList *device_list) +{ + cl_int status; + int i, j, k, device_num, total_devices_num, ret = 0; + int *devices_num; + cl_platform_id *platform_ids = NULL; + cl_device_id *device_ids = NULL; + AVOpenCLDeviceNode *device_node = NULL; + size_t platform_name_size = 0; + size_t device_name_size = 0; + status = clGetPlatformIDs(0, NULL, &device_list->platform_num); + if (status != CL_SUCCESS) { + av_log(&opencl_ctx, AV_LOG_ERROR, + "Could not get OpenCL platform ids: %s\n", av_opencl_errstr(status)); + return AVERROR_EXTERNAL; + } + platform_ids = av_mallocz_array(device_list->platform_num, sizeof(cl_platform_id)); + if (!platform_ids) + return AVERROR(ENOMEM); + status = clGetPlatformIDs(device_list->platform_num, platform_ids, NULL); + if (status != CL_SUCCESS) { + av_log(&opencl_ctx, AV_LOG_ERROR, + "Could not get OpenCL platform ids: %s\n", av_opencl_errstr(status)); + ret = AVERROR_EXTERNAL; + goto end; + } + device_list->platform_node = av_mallocz_array(device_list->platform_num, sizeof(AVOpenCLPlatformNode *)); + if (!device_list->platform_node) { + ret = AVERROR(ENOMEM); + goto end; + } + devices_num = av_mallocz(sizeof(int) * FF_ARRAY_ELEMS(device_type)); + if (!devices_num) { + ret = AVERROR(ENOMEM); + goto end; + } + for (i = 0; i < device_list->platform_num; i++) { + device_list->platform_node[i] = av_mallocz(sizeof(AVOpenCLPlatformNode)); + if (!device_list->platform_node[i]) { + ret = AVERROR(ENOMEM); + goto end; + } + device_list->platform_node[i]->platform_id = platform_ids[i]; + status = clGetPlatformInfo(platform_ids[i], CL_PLATFORM_VENDOR, + 0, NULL, &platform_name_size); + if (status != CL_SUCCESS) { + av_log(&opencl_ctx, AV_LOG_WARNING, + "Could not get size of platform name: %s\n", av_opencl_errstr(status)); + } else { + device_list->platform_node[i]->platform_name = av_malloc(platform_name_size * sizeof(char)); + if (!device_list->platform_node[i]->platform_name) { + av_log(&opencl_ctx, AV_LOG_WARNING, + "Could not allocate memory for device name: %s\n", av_opencl_errstr(status)); + } else { + status = clGetPlatformInfo(platform_ids[i], CL_PLATFORM_VENDOR, + platform_name_size * sizeof(char), + device_list->platform_node[i]->platform_name, NULL); + if (status != CL_SUCCESS) { + av_log(&opencl_ctx, AV_LOG_WARNING, + "Could not get platform name: %s\n", av_opencl_errstr(status)); + } + } + } + total_devices_num = 0; + for (j = 0; j < FF_ARRAY_ELEMS(device_type); j++) { + status = clGetDeviceIDs(device_list->platform_node[i]->platform_id, + device_type[j], 0, NULL, &devices_num[j]); + total_devices_num += devices_num[j]; + } + device_list->platform_node[i]->device_node = av_mallocz_array(total_devices_num, sizeof(AVOpenCLDeviceNode *)); + if (!device_list->platform_node[i]->device_node) { + ret = AVERROR(ENOMEM); + goto end; + } + for (j = 0; j < FF_ARRAY_ELEMS(device_type); j++) { + if (devices_num[j]) { + device_ids = av_mallocz_array(devices_num[j], sizeof(cl_device_id)); + if (!device_ids) { + ret = AVERROR(ENOMEM); + goto end; + } + status = clGetDeviceIDs(device_list->platform_node[i]->platform_id, device_type[j], + devices_num[j], device_ids, NULL); + if (status != CL_SUCCESS) { + av_log(&opencl_ctx, AV_LOG_WARNING, + "Could not get device ID: %s:\n", av_opencl_errstr(status)); + av_freep(&device_ids); + continue; + } + for (k = 0; k < devices_num[j]; k++) { + device_num = device_list->platform_node[i]->device_num; + device_list->platform_node[i]->device_node[device_num] = av_mallocz(sizeof(AVOpenCLDeviceNode)); + if (!device_list->platform_node[i]->device_node[device_num]) { + ret = AVERROR(ENOMEM); + goto end; + } + device_node = device_list->platform_node[i]->device_node[device_num]; + device_node->device_id = device_ids[k]; + device_node->device_type = device_type[j]; + status = clGetDeviceInfo(device_node->device_id, CL_DEVICE_NAME, + 0, NULL, &device_name_size); + if (status != CL_SUCCESS) { + av_log(&opencl_ctx, AV_LOG_WARNING, + "Could not get size of device name: %s\n", av_opencl_errstr(status)); + continue; + } + device_node->device_name = av_malloc(device_name_size * sizeof(char)); + if (!device_node->device_name) { + av_log(&opencl_ctx, AV_LOG_WARNING, + "Could not allocate memory for device name: %s\n", av_opencl_errstr(status)); + continue; + } + status = clGetDeviceInfo(device_node->device_id, CL_DEVICE_NAME, + device_name_size * sizeof(char), + device_node->device_name, NULL); + if (status != CL_SUCCESS) { + av_log(&opencl_ctx, AV_LOG_WARNING, + "Could not get device name: %s\n", av_opencl_errstr(status)); + continue; + } + device_list->platform_node[i]->device_num++; + } + av_freep(&device_ids); + } + } + } +end: + av_freep(&platform_ids); + av_freep(&devices_num); + av_freep(&device_ids); + if (ret < 0) + free_device_list(device_list); + return ret; +} + +int av_opencl_get_device_list(AVOpenCLDeviceList **device_list) +{ + int ret = 0; + *device_list = av_mallocz(sizeof(AVOpenCLDeviceList)); + if (!(*device_list)) { + av_log(&opencl_ctx, AV_LOG_ERROR, "Could not allocate opencl device list\n"); + return AVERROR(ENOMEM); + } + ret = get_device_list(*device_list); + if (ret < 0) { + av_log(&opencl_ctx, AV_LOG_ERROR, "Could not get device list from environment\n"); + free_device_list(*device_list); + av_freep(device_list); + return ret; + } + return ret; +} + +void av_opencl_free_device_list(AVOpenCLDeviceList **device_list) +{ + free_device_list(*device_list); + av_freep(device_list); +} + +static inline int init_opencl_mtx(void) +{ +#if HAVE_THREADS + if (!atomic_opencl_lock) { + int err; + pthread_mutex_t *tmp = av_malloc(sizeof(pthread_mutex_t)); + if (!tmp) + return AVERROR(ENOMEM); + if ((err = pthread_mutex_init(tmp, NULL))) { + av_free(tmp); + return AVERROR(err); + } + if (avpriv_atomic_ptr_cas((void * volatile *)&atomic_opencl_lock, NULL, tmp)) { + pthread_mutex_destroy(tmp); + av_free(tmp); + } + } +#endif + return 0; +} + +int av_opencl_set_option(const char *key, const char *val) +{ + int ret = init_opencl_mtx( ); + if (ret < 0) + return ret; + LOCK_OPENCL; + if (!opencl_ctx.opt_init_flag) { + av_opt_set_defaults(&opencl_ctx); + opencl_ctx.opt_init_flag = 1; + } + ret = av_opt_set(&opencl_ctx, key, val, 0); + UNLOCK_OPENCL; + return ret; +} + +int av_opencl_get_option(const char *key, uint8_t **out_val) +{ + int ret = 0; + LOCK_OPENCL; + ret = av_opt_get(&opencl_ctx, key, 0, out_val); + UNLOCK_OPENCL; + return ret; +} + +void av_opencl_free_option(void) +{ + /*FIXME: free openclutils context*/ + LOCK_OPENCL; + av_opt_free(&opencl_ctx); + UNLOCK_OPENCL; +} + +AVOpenCLExternalEnv *av_opencl_alloc_external_env(void) +{ + AVOpenCLExternalEnv *ext = av_mallocz(sizeof(AVOpenCLExternalEnv)); + if (!ext) { + av_log(&opencl_ctx, AV_LOG_ERROR, + "Could not malloc external opencl environment data space\n"); + } + return ext; +} + +void av_opencl_free_external_env(AVOpenCLExternalEnv **ext_opencl_env) +{ + av_freep(ext_opencl_env); +} + +int av_opencl_register_kernel_code(const char *kernel_code) +{ + int i, ret = init_opencl_mtx( ); + if (ret < 0) + return ret; + LOCK_OPENCL; + if (opencl_ctx.kernel_code_count >= MAX_KERNEL_CODE_NUM) { + av_log(&opencl_ctx, AV_LOG_ERROR, + "Could not register kernel code, maximum number of registered kernel code %d already reached\n", + MAX_KERNEL_CODE_NUM); + ret = AVERROR(EINVAL); + goto end; + } + for (i = 0; i < opencl_ctx.kernel_code_count; i++) { + if (opencl_ctx.kernel_code[i].kernel_string == kernel_code) { + av_log(&opencl_ctx, AV_LOG_WARNING, "Same kernel code has been registered\n"); + goto end; + } + } + opencl_ctx.kernel_code[opencl_ctx.kernel_code_count].kernel_string = kernel_code; + opencl_ctx.kernel_code[opencl_ctx.kernel_code_count].is_compiled = 0; + opencl_ctx.kernel_code_count++; +end: + UNLOCK_OPENCL; + return ret; +} + +cl_program av_opencl_compile(const char *program_name, const char *build_opts) +{ + int i; + cl_int status, build_status; + int kernel_code_idx = 0; + const char *kernel_source; + size_t kernel_code_len; + char* ptr = NULL; + cl_program program = NULL; + size_t log_size; + char *log = NULL; + + LOCK_OPENCL; + for (i = 0; i < opencl_ctx.kernel_code_count; i++) { + // identify a program using a unique name within the kernel source + ptr = av_stristr(opencl_ctx.kernel_code[i].kernel_string, program_name); + if (ptr && !opencl_ctx.kernel_code[i].is_compiled) { + kernel_source = opencl_ctx.kernel_code[i].kernel_string; + kernel_code_len = strlen(opencl_ctx.kernel_code[i].kernel_string); + kernel_code_idx = i; + break; + } + } + if (!kernel_source) { + av_log(&opencl_ctx, AV_LOG_ERROR, + "Unable to find OpenCL kernel source '%s'\n", program_name); + goto end; + } + + /* create a CL program from kernel source */ + program = clCreateProgramWithSource(opencl_ctx.context, 1, &kernel_source, &kernel_code_len, &status); + if(status != CL_SUCCESS) { + av_log(&opencl_ctx, AV_LOG_ERROR, + "Unable to create OpenCL program '%s': %s\n", program_name, av_opencl_errstr(status)); + program = NULL; + goto end; + } + + build_status = clBuildProgram(program, 1, &(opencl_ctx.device_id), build_opts, NULL, NULL); + status = clGetProgramBuildInfo(program, opencl_ctx.device_id, + CL_PROGRAM_BUILD_LOG, 0, NULL, &log_size); + if (status != CL_SUCCESS) { + av_log(&opencl_ctx, AV_LOG_WARNING, + "Failed to get compilation log: %s\n", + av_opencl_errstr(status)); + } else { + log = av_malloc(log_size); + if (log) { + status = clGetProgramBuildInfo(program, opencl_ctx.device_id, + CL_PROGRAM_BUILD_LOG, log_size, + log, NULL); + if (status != CL_SUCCESS) { + av_log(&opencl_ctx, AV_LOG_WARNING, + "Failed to get compilation log: %s\n", + av_opencl_errstr(status)); + } else { + int level = build_status == CL_SUCCESS ? AV_LOG_DEBUG : + AV_LOG_ERROR; + av_log(&opencl_ctx, level, "Compilation log:\n%s\n", log); + } + } + av_freep(&log); + } + if (build_status != CL_SUCCESS) { + av_log(&opencl_ctx, AV_LOG_ERROR, + "Compilation failed with OpenCL program '%s': %s\n", + program_name, av_opencl_errstr(build_status)); + program = NULL; + goto end; + } + + opencl_ctx.kernel_code[kernel_code_idx].is_compiled = 1; +end: + UNLOCK_OPENCL; + return program; +} + +cl_command_queue av_opencl_get_command_queue(void) +{ + return opencl_ctx.command_queue; +} + +static int init_opencl_env(OpenclContext *opencl_ctx, AVOpenCLExternalEnv *ext_opencl_env) +{ + cl_int status; + cl_context_properties cps[3]; + int i, ret = 0; + AVOpenCLDeviceNode *device_node = NULL; + + if (ext_opencl_env) { + if (opencl_ctx->is_user_created) + return 0; + opencl_ctx->platform_id = ext_opencl_env->platform_id; + opencl_ctx->is_user_created = 1; + opencl_ctx->command_queue = ext_opencl_env->command_queue; + opencl_ctx->context = ext_opencl_env->context; + opencl_ctx->device_id = ext_opencl_env->device_id; + opencl_ctx->device_type = ext_opencl_env->device_type; + } else { + if (!opencl_ctx->is_user_created) { + if (!opencl_ctx->device_list.platform_num) { + ret = get_device_list(&opencl_ctx->device_list); + if (ret < 0) { + return ret; + } + } + if (opencl_ctx->platform_idx >= 0) { + if (opencl_ctx->device_list.platform_num < opencl_ctx->platform_idx + 1) { + av_log(opencl_ctx, AV_LOG_ERROR, "User set platform index not exist\n"); + return AVERROR(EINVAL); + } + if (!opencl_ctx->device_list.platform_node[opencl_ctx->platform_idx]->device_num) { + av_log(opencl_ctx, AV_LOG_ERROR, "No devices in user specific platform with index %d\n", + opencl_ctx->platform_idx); + return AVERROR(EINVAL); + } + opencl_ctx->platform_id = opencl_ctx->device_list.platform_node[opencl_ctx->platform_idx]->platform_id; + } else { + /* get a usable platform by default*/ + for (i = 0; i < opencl_ctx->device_list.platform_num; i++) { + if (opencl_ctx->device_list.platform_node[i]->device_num) { + opencl_ctx->platform_id = opencl_ctx->device_list.platform_node[i]->platform_id; + opencl_ctx->platform_idx = i; + break; + } + } + } + if (!opencl_ctx->platform_id) { + av_log(opencl_ctx, AV_LOG_ERROR, "Could not get OpenCL platforms\n"); + return AVERROR_EXTERNAL; + } + /* get a usable device*/ + if (opencl_ctx->device_idx >= 0) { + if (opencl_ctx->device_list.platform_node[opencl_ctx->platform_idx]->device_num < opencl_ctx->device_idx + 1) { + av_log(opencl_ctx, AV_LOG_ERROR, + "Could not get OpenCL device idx %d in the user set platform\n", opencl_ctx->platform_idx); + return AVERROR(EINVAL); + } + } else { + opencl_ctx->device_idx = 0; + } + + device_node = opencl_ctx->device_list.platform_node[opencl_ctx->platform_idx]->device_node[opencl_ctx->device_idx]; + opencl_ctx->device_id = device_node->device_id; + opencl_ctx->device_type = device_node->device_type; + + /* + * Use available platform. + */ + av_log(opencl_ctx, AV_LOG_VERBOSE, "Platform Name: %s, Device Name: %s\n", + opencl_ctx->device_list.platform_node[opencl_ctx->platform_idx]->platform_name, + device_node->device_name); + cps[0] = CL_CONTEXT_PLATFORM; + cps[1] = (cl_context_properties)opencl_ctx->platform_id; + cps[2] = 0; + + opencl_ctx->context = clCreateContextFromType(cps, opencl_ctx->device_type, + NULL, NULL, &status); + if (status != CL_SUCCESS) { + av_log(opencl_ctx, AV_LOG_ERROR, + "Could not get OpenCL context from device type: %s\n", av_opencl_errstr(status)); + return AVERROR_EXTERNAL; + } + opencl_ctx->command_queue = clCreateCommandQueue(opencl_ctx->context, opencl_ctx->device_id, + 0, &status); + if (status != CL_SUCCESS) { + av_log(opencl_ctx, AV_LOG_ERROR, + "Could not create OpenCL command queue: %s\n", av_opencl_errstr(status)); + return AVERROR_EXTERNAL; + } + } + } + return ret; +} + +int av_opencl_init(AVOpenCLExternalEnv *ext_opencl_env) +{ + int ret = init_opencl_mtx( ); + if (ret < 0) + return ret; + LOCK_OPENCL; + if (!opencl_ctx.init_count) { + if (!opencl_ctx.opt_init_flag) { + av_opt_set_defaults(&opencl_ctx); + opencl_ctx.opt_init_flag = 1; + } + ret = init_opencl_env(&opencl_ctx, ext_opencl_env); + if (ret < 0) + goto end; + if (opencl_ctx.kernel_code_count <= 0) { + av_log(&opencl_ctx, AV_LOG_ERROR, + "No kernel code is registered, compile kernel file failed\n"); + ret = AVERROR(EINVAL); + goto end; + } + } + opencl_ctx.init_count++; +end: + UNLOCK_OPENCL; + return ret; +} + +void av_opencl_uninit(void) +{ + int i; + cl_int status; + LOCK_OPENCL; + opencl_ctx.init_count--; + if (opencl_ctx.is_user_created) + goto end; + if (opencl_ctx.init_count > 0) + goto end; + if (opencl_ctx.command_queue) { + status = clReleaseCommandQueue(opencl_ctx.command_queue); + if (status != CL_SUCCESS) { + av_log(&opencl_ctx, AV_LOG_ERROR, + "Could not release OpenCL command queue: %s\n", av_opencl_errstr(status)); + } + opencl_ctx.command_queue = NULL; + } + if (opencl_ctx.context) { + status = clReleaseContext(opencl_ctx.context); + if (status != CL_SUCCESS) { + av_log(&opencl_ctx, AV_LOG_ERROR, + "Could not release OpenCL context: %s\n", av_opencl_errstr(status)); + } + opencl_ctx.context = NULL; + } + for (i = 0; i < opencl_ctx.kernel_code_count; i++) { + opencl_ctx.kernel_code[i].is_compiled = 0; + } + free_device_list(&opencl_ctx.device_list); +end: + if (opencl_ctx.init_count <= 0) + av_opt_free(&opencl_ctx); //FIXME: free openclutils context + UNLOCK_OPENCL; +} + +int av_opencl_buffer_create(cl_mem *cl_buf, size_t cl_buf_size, int flags, void *host_ptr) +{ + cl_int status; + *cl_buf = clCreateBuffer(opencl_ctx.context, flags, cl_buf_size, host_ptr, &status); + if (status != CL_SUCCESS) { + av_log(&opencl_ctx, AV_LOG_ERROR, "Could not create OpenCL buffer: %s\n", av_opencl_errstr(status)); + return AVERROR_EXTERNAL; + } + return 0; +} + +void av_opencl_buffer_release(cl_mem *cl_buf) +{ + cl_int status = 0; + if (!cl_buf) + return; + status = clReleaseMemObject(*cl_buf); + if (status != CL_SUCCESS) { + av_log(&opencl_ctx, AV_LOG_ERROR, + "Could not release OpenCL buffer: %s\n", av_opencl_errstr(status)); + } + memset(cl_buf, 0, sizeof(*cl_buf)); +} + +int av_opencl_buffer_write(cl_mem dst_cl_buf, uint8_t *src_buf, size_t buf_size) +{ + cl_int status; + void *mapped = clEnqueueMapBuffer(opencl_ctx.command_queue, dst_cl_buf, + CL_TRUE, CL_MAP_WRITE, 0, sizeof(uint8_t) * buf_size, + 0, NULL, NULL, &status); + + if (status != CL_SUCCESS) { + av_log(&opencl_ctx, AV_LOG_ERROR, + "Could not map OpenCL buffer: %s\n", av_opencl_errstr(status)); + return AVERROR_EXTERNAL; + } + memcpy(mapped, src_buf, buf_size); + + status = clEnqueueUnmapMemObject(opencl_ctx.command_queue, dst_cl_buf, mapped, 0, NULL, NULL); + if (status != CL_SUCCESS) { + av_log(&opencl_ctx, AV_LOG_ERROR, + "Could not unmap OpenCL buffer: %s\n", av_opencl_errstr(status)); + return AVERROR_EXTERNAL; + } + return 0; +} + +int av_opencl_buffer_read(uint8_t *dst_buf, cl_mem src_cl_buf, size_t buf_size) +{ + cl_int status; + void *mapped = clEnqueueMapBuffer(opencl_ctx.command_queue, src_cl_buf, + CL_TRUE, CL_MAP_READ, 0, buf_size, + 0, NULL, NULL, &status); + + if (status != CL_SUCCESS) { + av_log(&opencl_ctx, AV_LOG_ERROR, + "Could not map OpenCL buffer: %s\n", av_opencl_errstr(status)); + return AVERROR_EXTERNAL; + } + memcpy(dst_buf, mapped, buf_size); + + status = clEnqueueUnmapMemObject(opencl_ctx.command_queue, src_cl_buf, mapped, 0, NULL, NULL); + if (status != CL_SUCCESS) { + av_log(&opencl_ctx, AV_LOG_ERROR, + "Could not unmap OpenCL buffer: %s\n", av_opencl_errstr(status)); + return AVERROR_EXTERNAL; + } + return 0; +} + +int av_opencl_buffer_write_image(cl_mem dst_cl_buf, size_t cl_buffer_size, int dst_cl_offset, + uint8_t **src_data, int *plane_size, int plane_num) +{ + int i, buffer_size = 0; + uint8_t *temp; + cl_int status; + void *mapped; + if ((unsigned int)plane_num > 8) { + return AVERROR(EINVAL); + } + for (i = 0;i < plane_num;i++) { + buffer_size += plane_size[i]; + } + if (buffer_size > cl_buffer_size) { + av_log(&opencl_ctx, AV_LOG_ERROR, + "Cannot write image to OpenCL buffer: buffer too small\n"); + return AVERROR(EINVAL); + } + mapped = clEnqueueMapBuffer(opencl_ctx.command_queue, dst_cl_buf, + CL_TRUE, CL_MAP_WRITE, 0, buffer_size + dst_cl_offset, + 0, NULL, NULL, &status); + if (status != CL_SUCCESS) { + av_log(&opencl_ctx, AV_LOG_ERROR, + "Could not map OpenCL buffer: %s\n", av_opencl_errstr(status)); + return AVERROR_EXTERNAL; + } + temp = mapped; + temp += dst_cl_offset; + for (i = 0; i < plane_num; i++) { + memcpy(temp, src_data[i], plane_size[i]); + temp += plane_size[i]; + } + status = clEnqueueUnmapMemObject(opencl_ctx.command_queue, dst_cl_buf, mapped, 0, NULL, NULL); + if (status != CL_SUCCESS) { + av_log(&opencl_ctx, AV_LOG_ERROR, + "Could not unmap OpenCL buffer: %s\n", av_opencl_errstr(status)); + return AVERROR_EXTERNAL; + } + return 0; +} + +int av_opencl_buffer_read_image(uint8_t **dst_data, int *plane_size, int plane_num, + cl_mem src_cl_buf, size_t cl_buffer_size) +{ + int i,buffer_size = 0,ret = 0; + uint8_t *temp; + void *mapped; + cl_int status; + if ((unsigned int)plane_num > 8) { + return AVERROR(EINVAL); + } + for (i = 0; i < plane_num; i++) { + buffer_size += plane_size[i]; + } + if (buffer_size > cl_buffer_size) { + av_log(&opencl_ctx, AV_LOG_ERROR, + "Cannot write image to CPU buffer: OpenCL buffer too small\n"); + return AVERROR(EINVAL); + } + mapped = clEnqueueMapBuffer(opencl_ctx.command_queue, src_cl_buf, + CL_TRUE, CL_MAP_READ, 0, buffer_size, + 0, NULL, NULL, &status); + + if (status != CL_SUCCESS) { + av_log(&opencl_ctx, AV_LOG_ERROR, + "Could not map OpenCL buffer: %s\n", av_opencl_errstr(status)); + return AVERROR_EXTERNAL; + } + temp = mapped; + if (ret >= 0) { + for (i = 0; i < plane_num; i++) { + memcpy(dst_data[i], temp, plane_size[i]); + temp += plane_size[i]; + } + } + status = clEnqueueUnmapMemObject(opencl_ctx.command_queue, src_cl_buf, mapped, 0, NULL, NULL); + if (status != CL_SUCCESS) { + av_log(&opencl_ctx, AV_LOG_ERROR, + "Could not unmap OpenCL buffer: %s\n", av_opencl_errstr(status)); + return AVERROR_EXTERNAL; + } + return 0; +} + +int64_t av_opencl_benchmark(AVOpenCLDeviceNode *device_node, cl_platform_id platform, + int64_t (*benchmark)(AVOpenCLExternalEnv *ext_opencl_env)) +{ + int64_t ret = 0; + cl_int status; + cl_context_properties cps[3]; + AVOpenCLExternalEnv *ext_opencl_env = NULL; + + ext_opencl_env = av_opencl_alloc_external_env(); + ext_opencl_env->device_id = device_node->device_id; + ext_opencl_env->device_type = device_node->device_type; + av_log(&opencl_ctx, AV_LOG_VERBOSE, "Performing test on OpenCL device %s\n", + device_node->device_name); + + cps[0] = CL_CONTEXT_PLATFORM; + cps[1] = (cl_context_properties)platform; + cps[2] = 0; + ext_opencl_env->context = clCreateContextFromType(cps, ext_opencl_env->device_type, + NULL, NULL, &status); + if (status != CL_SUCCESS || !ext_opencl_env->context) { + ret = AVERROR_EXTERNAL; + goto end; + } + ext_opencl_env->command_queue = clCreateCommandQueue(ext_opencl_env->context, + ext_opencl_env->device_id, 0, &status); + if (status != CL_SUCCESS || !ext_opencl_env->command_queue) { + ret = AVERROR_EXTERNAL; + goto end; + } + ret = benchmark(ext_opencl_env); + if (ret < 0) + av_log(&opencl_ctx, AV_LOG_ERROR, "Benchmark failed with OpenCL device %s\n", + device_node->device_name); +end: + if (ext_opencl_env->command_queue) + clReleaseCommandQueue(ext_opencl_env->command_queue); + if (ext_opencl_env->context) + clReleaseContext(ext_opencl_env->context); + av_opencl_free_external_env(&ext_opencl_env); + return ret; +} diff --git a/libavutil/opencl.h b/libavutil/opencl.h new file mode 100644 index 0000000000..b709927c42 --- /dev/null +++ b/libavutil/opencl.h @@ -0,0 +1,292 @@ +/* + * Copyright (C) 2012 Peng Gao <peng@multicorewareinc.com> + * Copyright (C) 2012 Li Cao <li@multicorewareinc.com> + * Copyright (C) 2012 Wei Gao <weigao@multicorewareinc.com> + * Copyright (C) 2013 Lenny Wang <lwanghpc@gmail.com> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +/** + * @file + * OpenCL wrapper + * + * This interface is considered still experimental and its API and ABI may + * change without prior notice. + */ + +#ifndef AVUTIL_OPENCL_H +#define AVUTIL_OPENCL_H + +#define CL_USE_DEPRECATED_OPENCL_1_2_APIS 1 +#ifdef __APPLE__ +#include <OpenCL/cl.h> +#else +#include <CL/cl.h> +#endif +#include <stdint.h> +#include "dict.h" + +#include "libavutil/version.h" + +#define AV_OPENCL_KERNEL( ... )# __VA_ARGS__ + +typedef struct { + int device_type; + char *device_name; + cl_device_id device_id; +} AVOpenCLDeviceNode; + +typedef struct { + cl_platform_id platform_id; + char *platform_name; + int device_num; + AVOpenCLDeviceNode **device_node; +} AVOpenCLPlatformNode; + +typedef struct { + int platform_num; + AVOpenCLPlatformNode **platform_node; +} AVOpenCLDeviceList; + +typedef struct { + cl_platform_id platform_id; + cl_device_type device_type; + cl_context context; + cl_device_id device_id; + cl_command_queue command_queue; + char *platform_name; +} AVOpenCLExternalEnv; + +/** + * Get OpenCL device list. + * + * It must be freed with av_opencl_free_device_list(). + * + * @param device_list pointer to OpenCL environment device list, + * should be released by av_opencl_free_device_list() + * + * @return >=0 on success, a negative error code in case of failure + */ +int av_opencl_get_device_list(AVOpenCLDeviceList **device_list); + +/** + * Free OpenCL device list. + * + * @param device_list pointer to OpenCL environment device list + * created by av_opencl_get_device_list() + */ +void av_opencl_free_device_list(AVOpenCLDeviceList **device_list); + +/** + * Set option in the global OpenCL context. + * + * This options affect the operation performed by the next + * av_opencl_init() operation. + * + * The currently accepted options are: + * - platform: set index of platform in device list + * - device: set index of device in device list + * + * See reference "OpenCL Specification Version: 1.2 chapter 5.6.4". + * + * @param key option key + * @param val option value + * @return >=0 on success, a negative error code in case of failure + * @see av_opencl_get_option() + */ +int av_opencl_set_option(const char *key, const char *val); + +/** + * Get option value from the global OpenCL context. + * + * @param key option key + * @param out_val pointer to location where option value will be + * written, must be freed with av_freep() + * @return >=0 on success, a negative error code in case of failure + * @see av_opencl_set_option() + */ +int av_opencl_get_option(const char *key, uint8_t **out_val); + +/** + * Free option values of the global OpenCL context. + * + */ +void av_opencl_free_option(void); + +/** + * Allocate OpenCL external environment. + * + * It must be freed with av_opencl_free_external_env(). + * + * @return pointer to allocated OpenCL external environment + */ +AVOpenCLExternalEnv *av_opencl_alloc_external_env(void); + +/** + * Free OpenCL external environment. + * + * @param ext_opencl_env pointer to OpenCL external environment + * created by av_opencl_alloc_external_env() + */ +void av_opencl_free_external_env(AVOpenCLExternalEnv **ext_opencl_env); + +/** + * Get OpenCL error string. + * + * @param status OpenCL error code + * @return OpenCL error string + */ +const char *av_opencl_errstr(cl_int status); + +/** + * Register kernel code. + * + * The registered kernel code is stored in a global context, and compiled + * in the runtime environment when av_opencl_init() is called. + * + * @param kernel_code kernel code to be compiled in the OpenCL runtime environment + * @return >=0 on success, a negative error code in case of failure + */ +int av_opencl_register_kernel_code(const char *kernel_code); + +/** + * Initialize the run time OpenCL environment + * + * @param ext_opencl_env external OpenCL environment, created by an + * application program, ignored if set to NULL + * @return >=0 on success, a negative error code in case of failure + */ +int av_opencl_init(AVOpenCLExternalEnv *ext_opencl_env); + +/** + * compile specific OpenCL kernel source + * + * @param program_name pointer to a program name used for identification + * @param build_opts pointer to a string that describes the preprocessor + * build options to be used for building the program + * @return a cl_program object + */ +cl_program av_opencl_compile(const char *program_name, const char* build_opts); + +/** + * get OpenCL command queue + * + * @return a cl_command_queue object + */ +cl_command_queue av_opencl_get_command_queue(void); + +/** + * Create OpenCL buffer. + * + * The buffer is used to save the data used or created by an OpenCL + * kernel. + * The created buffer must be released with av_opencl_buffer_release(). + * + * See clCreateBuffer() function reference for more information about + * the parameters. + * + * @param cl_buf pointer to OpenCL buffer + * @param cl_buf_size size in bytes of the OpenCL buffer to create + * @param flags flags used to control buffer attributes + * @param host_ptr host pointer of the OpenCL buffer + * @return >=0 on success, a negative error code in case of failure + */ +int av_opencl_buffer_create(cl_mem *cl_buf, size_t cl_buf_size, int flags, void *host_ptr); + +/** + * Write OpenCL buffer with data from src_buf. + * + * @param dst_cl_buf pointer to OpenCL destination buffer + * @param src_buf pointer to source buffer + * @param buf_size size in bytes of the source and destination buffers + * @return >=0 on success, a negative error code in case of failure + */ +int av_opencl_buffer_write(cl_mem dst_cl_buf, uint8_t *src_buf, size_t buf_size); + +/** + * Read data from OpenCL buffer to memory buffer. + * + * @param dst_buf pointer to destination buffer (CPU memory) + * @param src_cl_buf pointer to source OpenCL buffer + * @param buf_size size in bytes of the source and destination buffers + * @return >=0 on success, a negative error code in case of failure + */ +int av_opencl_buffer_read(uint8_t *dst_buf, cl_mem src_cl_buf, size_t buf_size); + +/** + * Write image data from memory to OpenCL buffer. + * + * The source must be an array of pointers to image plane buffers. + * + * @param dst_cl_buf pointer to destination OpenCL buffer + * @param dst_cl_buf_size size in bytes of OpenCL buffer + * @param dst_cl_buf_offset the offset of the OpenCL buffer start position + * @param src_data array of pointers to source plane buffers + * @param src_plane_sizes array of sizes in bytes of the source plane buffers + * @param src_plane_num number of source image planes + * @return >=0 on success, a negative error code in case of failure + */ +int av_opencl_buffer_write_image(cl_mem dst_cl_buf, size_t cl_buffer_size, int dst_cl_offset, + uint8_t **src_data, int *plane_size, int plane_num); + +/** + * Read image data from OpenCL buffer. + * + * @param dst_data array of pointers to destination plane buffers + * @param dst_plane_sizes array of pointers to destination plane buffers + * @param dst_plane_num number of destination image planes + * @param src_cl_buf pointer to source OpenCL buffer + * @param src_cl_buf_size size in bytes of OpenCL buffer + * @return >=0 on success, a negative error code in case of failure + */ +int av_opencl_buffer_read_image(uint8_t **dst_data, int *plane_size, int plane_num, + cl_mem src_cl_buf, size_t cl_buffer_size); + +/** + * Release OpenCL buffer. + * + * @param cl_buf pointer to OpenCL buffer to release, which was + * previously filled with av_opencl_buffer_create() + */ +void av_opencl_buffer_release(cl_mem *cl_buf); + +/** + * Release OpenCL environment. + * + * The OpenCL environment is effectively released only if all the created + * kernels had been released with av_opencl_release_kernel(). + */ +void av_opencl_uninit(void); + +/** + * Benchmark an OpenCL device with a user defined callback function. This function + * sets up an external OpenCL environment including context and command queue on + * the device then tears it down in the end. The callback function should perform + * the rest of the work. + * + * @param device pointer to the OpenCL device to be used + * @param platform cl_platform_id handle to which the device belongs to + * @param benchmark callback function to perform the benchmark, return a + * negative value in case of failure + * @return the score passed from the callback function, a negative error code in case + * of failure + */ +int64_t av_opencl_benchmark(AVOpenCLDeviceNode *device, cl_platform_id platform, + int64_t (*benchmark)(AVOpenCLExternalEnv *ext_opencl_env)); + +#endif /* AVUTIL_OPENCL_H */ diff --git a/libavutil/opencl_internal.c b/libavutil/opencl_internal.c new file mode 100644 index 0000000000..bdb41936fa --- /dev/null +++ b/libavutil/opencl_internal.c @@ -0,0 +1,59 @@ +/* + * Copyright (C) 2012 Peng Gao <peng@multicorewareinc.com> + * Copyright (C) 2012 Li Cao <li@multicorewareinc.com> + * Copyright (C) 2012 Wei Gao <weigao@multicorewareinc.com> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "opencl_internal.h" +#include "libavutil/log.h" + +int avpriv_opencl_set_parameter(FFOpenclParam *opencl_param, ...) +{ + int ret = 0; + va_list arg_ptr; + void *param; + size_t param_size; + cl_int status; + if (!opencl_param->kernel) { + av_log(opencl_param->ctx, AV_LOG_ERROR, "OpenCL kernel must be set\n"); + return AVERROR(EINVAL); + } + va_start(arg_ptr, opencl_param); + do { + param = va_arg(arg_ptr, void *); + if (!param) + break; + param_size = va_arg(arg_ptr, size_t); + if (!param_size) { + av_log(opencl_param->ctx, AV_LOG_ERROR, "Parameter size must not be 0\n"); + ret = AVERROR(EINVAL); + goto end; + } + status = clSetKernelArg(opencl_param->kernel, opencl_param->param_num, param_size, param); + if (status != CL_SUCCESS) { + av_log(opencl_param->ctx, AV_LOG_ERROR, "Cannot set kernel argument: %s\n", av_opencl_errstr(status)); + ret = AVERROR_EXTERNAL; + goto end; + } + opencl_param->param_num++; + } while (param && param_size); +end: + va_end(arg_ptr); + return ret; +} diff --git a/libavutil/opencl_internal.h b/libavutil/opencl_internal.h new file mode 100644 index 0000000000..5cabb7b3ee --- /dev/null +++ b/libavutil/opencl_internal.h @@ -0,0 +1,40 @@ +/* + * Copyright (C) 2012 Peng Gao <peng@multicorewareinc.com> + * Copyright (C) 2012 Li Cao <li@multicorewareinc.com> + * Copyright (C) 2012 Wei Gao <weigao@multicorewareinc.com> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#ifndef AVUTIL_OPENCL_INTERNAL_H +#define AVUTIL_OPENCL_INTERNAL_H + +#include "attributes.h" +#include "opencl.h" + +#define FF_OPENCL_PARAM_INFO(a) ((void*)(&(a))), (sizeof(a)) + +typedef struct { + cl_kernel kernel; + int param_num; + void *ctx; +} FFOpenclParam; + +av_warn_unused_result +int avpriv_opencl_set_parameter(FFOpenclParam *opencl_param, ...); + +#endif /* AVUTIL_OPENCL_INTERNAL_H */ diff --git a/libavutil/opt-test.c b/libavutil/opt-test.c index 14a144cb86..1f813d69a5 100644 --- a/libavutil/opt-test.c +++ b/libavutil/opt-test.c @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -20,11 +20,13 @@ #include <stdio.h> #include "common.h" +#include "channel_layout.h" #include "error.h" #include "log.h" #include "mem.h" #include "rational.h" #include "opt.h" +#include "pixdesc.h" typedef struct TestContext { const AVClass *class; @@ -33,6 +35,26 @@ typedef struct TestContext { char *string; int flags; AVRational rational; + AVRational video_rate; + int w, h; + enum AVPixelFormat pix_fmt; + enum AVSampleFormat sample_fmt; + int64_t duration; + uint8_t color[4]; + int64_t channel_layout; + void *binary; + int binary_size; + void *binary1; + int binary_size1; + void *binary2; + int binary_size2; + int64_t num64; + float flt; + double dbl; + char *escape; + int bool1; + int bool2; + int bool3; } TestContext; #define OFFSET(x) offsetof(TestContext, x) @@ -41,15 +63,32 @@ typedef struct TestContext { #define TEST_FLAG_LAME 02 #define TEST_FLAG_MU 04 -static const AVOption test_options[] = { - { "num", "set num", OFFSET(num), AV_OPT_TYPE_INT, { .i64 = 0 }, 0, 100 }, - { "toggle", "set toggle", OFFSET(toggle), AV_OPT_TYPE_INT, { .i64 = 0 }, 0, 1 }, - { "rational", "set rational", OFFSET(rational), AV_OPT_TYPE_RATIONAL, { .dbl = 0 }, 0, 10 }, - { "string", "set string", OFFSET(string), AV_OPT_TYPE_STRING, { 0 }, CHAR_MIN, CHAR_MAX }, - { "flags", "set flags", OFFSET(flags), AV_OPT_TYPE_FLAGS, { .i64 = 0 }, 0, INT_MAX, 0, "flags"}, - { "cool", "set cool flag ", 0, AV_OPT_TYPE_CONST, { .i64 = TEST_FLAG_COOL }, INT_MIN, INT_MAX, 0, "flags"}, - { "lame", "set lame flag ", 0, AV_OPT_TYPE_CONST, { .i64 = TEST_FLAG_LAME }, INT_MIN, INT_MAX, 0, "flags"}, - { "mu", "set mu flag ", 0, AV_OPT_TYPE_CONST, { .i64 = TEST_FLAG_MU }, INT_MIN, INT_MAX, 0, "flags"}, +static const AVOption test_options[]= { + {"num", "set num", OFFSET(num), AV_OPT_TYPE_INT, { .i64 = 0 }, 0, 100, 1 }, + {"toggle", "set toggle", OFFSET(toggle), AV_OPT_TYPE_INT, { .i64 = 1 }, 0, 1, 1 }, + {"rational", "set rational", OFFSET(rational), AV_OPT_TYPE_RATIONAL, { .dbl = 1 }, 0, 10, 1 }, + {"string", "set string", OFFSET(string), AV_OPT_TYPE_STRING, { .str = "default" }, CHAR_MIN, CHAR_MAX, 1 }, + {"escape", "set escape str", OFFSET(escape), AV_OPT_TYPE_STRING, { .str = "\\=," }, CHAR_MIN, CHAR_MAX, 1 }, + {"flags", "set flags", OFFSET(flags), AV_OPT_TYPE_FLAGS, { .i64 = 1 }, 0, INT_MAX, 1, "flags" }, + {"cool", "set cool flag", 0, AV_OPT_TYPE_CONST, { .i64 = TEST_FLAG_COOL }, INT_MIN, INT_MAX, 1, "flags" }, + {"lame", "set lame flag", 0, AV_OPT_TYPE_CONST, { .i64 = TEST_FLAG_LAME }, INT_MIN, INT_MAX, 1, "flags" }, + {"mu", "set mu flag", 0, AV_OPT_TYPE_CONST, { .i64 = TEST_FLAG_MU }, INT_MIN, INT_MAX, 1, "flags" }, + {"size", "set size", OFFSET(w), AV_OPT_TYPE_IMAGE_SIZE, { .str="200x300" }, 0, 0, 1 }, + {"pix_fmt", "set pixfmt", OFFSET(pix_fmt), AV_OPT_TYPE_PIXEL_FMT, { .i64 = AV_PIX_FMT_0BGR }, -1, INT_MAX, 1 }, + {"sample_fmt", "set samplefmt", OFFSET(sample_fmt), AV_OPT_TYPE_SAMPLE_FMT, { .i64 = AV_SAMPLE_FMT_S16 }, -1, INT_MAX, 1 }, + {"video_rate", "set videorate", OFFSET(video_rate), AV_OPT_TYPE_VIDEO_RATE, { .str = "25" }, 0, 0, 1 }, + {"duration", "set duration", OFFSET(duration), AV_OPT_TYPE_DURATION, { .i64 = 1000 }, 0, INT64_MAX, 1 }, + {"color", "set color", OFFSET(color), AV_OPT_TYPE_COLOR, { .str = "pink" }, 0, 0, 1 }, + {"cl", "set channel layout", OFFSET(channel_layout), AV_OPT_TYPE_CHANNEL_LAYOUT, { .i64 = AV_CH_LAYOUT_HEXAGONAL }, 0, INT64_MAX, 1 }, + {"bin", "set binary value", OFFSET(binary), AV_OPT_TYPE_BINARY, { .str="62696e00" }, 0, 0, 1 }, + {"bin1", "set binary value", OFFSET(binary1), AV_OPT_TYPE_BINARY, { .str=NULL }, 0, 0, 1 }, + {"bin2", "set binary value", OFFSET(binary2), AV_OPT_TYPE_BINARY, { .str="" }, 0, 0, 1 }, + {"num64", "set num 64bit", OFFSET(num64), AV_OPT_TYPE_INT64, { .i64 = 1 }, 0, 100, 1 }, + {"flt", "set float", OFFSET(flt), AV_OPT_TYPE_FLOAT, { .dbl = 1.0 / 3 }, 0, 100, 1 }, + {"dbl", "set double", OFFSET(dbl), AV_OPT_TYPE_DOUBLE, { .dbl = 1.0 / 3 }, 0, 100, 1 }, + {"bool1", "set boolean value", OFFSET(bool1), AV_OPT_TYPE_BOOL, { .i64 = -1 }, -1, 1, 1 }, + {"bool2", "set boolean value", OFFSET(bool2), AV_OPT_TYPE_BOOL, { .i64 = 1 }, -1, 1, 1 }, + {"bool3", "set boolean value", OFFSET(bool3), AV_OPT_TYPE_BOOL, { .i64 = 0 }, 0, 1, 1 }, { NULL }, }; @@ -64,46 +103,209 @@ static const AVClass test_class = { test_options }; +static void log_callback_help(void *ptr, int level, const char *fmt, va_list vl) +{ + vfprintf(stdout, fmt, vl); +} + int main(void) { int i; - TestContext test_ctx = { .class = &test_class }; - const char *options[] = { - "", - ":", - "=", - "foo=:", - ":=foo", - "=foo", - "foo=", - "foo", - "foo=val", - "foo==val", - "toggle=:", - "string=:", - "toggle=1 : foo", - "toggle=100", - "toggle==1", - "flags=+mu-lame : num=42: toggle=0", - "num=42 : string=blahblah", - "rational=0 : rational=1/2 : rational=1/-1", - "rational=-1/0", - }; + + av_log_set_level(AV_LOG_DEBUG); + av_log_set_callback(log_callback_help); + + printf("Testing default values\n"); + { + TestContext test_ctx = { 0 }; + test_ctx.class = &test_class; + av_opt_set_defaults(&test_ctx); + + printf("num=%d\n", test_ctx.num); + printf("toggle=%d\n", test_ctx.toggle); + printf("string=%s\n", test_ctx.string); + printf("escape=%s\n", test_ctx.escape); + printf("flags=%d\n", test_ctx.flags); + printf("rational=%d/%d\n", test_ctx.rational.num, test_ctx.rational.den); + printf("video_rate=%d/%d\n", test_ctx.video_rate.num, test_ctx.video_rate.den); + printf("width=%d height=%d\n", test_ctx.w, test_ctx.h); + printf("pix_fmt=%s\n", av_get_pix_fmt_name(test_ctx.pix_fmt)); + printf("sample_fmt=%s\n", av_get_sample_fmt_name(test_ctx.sample_fmt)); + printf("duration=%"PRId64"\n", test_ctx.duration); + printf("color=%d %d %d %d\n", test_ctx.color[0], test_ctx.color[1], test_ctx.color[2], test_ctx.color[3]); + printf("channel_layout=%"PRId64"=%"PRId64"\n", test_ctx.channel_layout, (int64_t)AV_CH_LAYOUT_HEXAGONAL); + if (test_ctx.binary) + printf("binary=%x %x %x %x\n", ((uint8_t*)test_ctx.binary)[0], ((uint8_t*)test_ctx.binary)[1], ((uint8_t*)test_ctx.binary)[2], ((uint8_t*)test_ctx.binary)[3]); + printf("binary_size=%d\n", test_ctx.binary_size); + printf("num64=%"PRId64"\n", test_ctx.num64); + printf("flt=%.6f\n", test_ctx.flt); + printf("dbl=%.6f\n", test_ctx.dbl); + + av_opt_show2(&test_ctx, NULL, -1, 0); + + av_opt_free(&test_ctx); + } + + printf("\nTesting av_opt_is_set_to_default()\n"); + { + int ret; + TestContext test_ctx = { 0 }; + const AVOption *o = NULL; + test_ctx.class = &test_class; + + av_log_set_level(AV_LOG_QUIET); + + while (o = av_opt_next(&test_ctx, o)) { + ret = av_opt_is_set_to_default_by_name(&test_ctx, o->name, 0); + printf("name:%10s default:%d error:%s\n", o->name, !!ret, ret < 0 ? av_err2str(ret) : ""); + } + av_opt_set_defaults(&test_ctx); + while (o = av_opt_next(&test_ctx, o)) { + ret = av_opt_is_set_to_default_by_name(&test_ctx, o->name, 0); + printf("name:%10s default:%d error:%s\n", o->name, !!ret, ret < 0 ? av_err2str(ret) : ""); + } + av_opt_free(&test_ctx); + } + + printf("\nTest av_opt_serialize()\n"); + { + TestContext test_ctx = { 0 }; + char *buf; + test_ctx.class = &test_class; + + av_log_set_level(AV_LOG_QUIET); + + av_opt_set_defaults(&test_ctx); + if (av_opt_serialize(&test_ctx, 0, 0, &buf, '=', ',') >= 0) { + printf("%s\n", buf); + av_opt_free(&test_ctx); + memset(&test_ctx, 0, sizeof(test_ctx)); + test_ctx.class = &test_class; + av_set_options_string(&test_ctx, buf, "=", ","); + av_free(buf); + if (av_opt_serialize(&test_ctx, 0, 0, &buf, '=', ',') >= 0) { + printf("%s\n", buf); + av_free(buf); + } + } + av_opt_free(&test_ctx); + } printf("\nTesting av_set_options_string()\n"); + { + TestContext test_ctx = { 0 }; + static const char * const options[] = { + "", + ":", + "=", + "foo=:", + ":=foo", + "=foo", + "foo=", + "foo", + "foo=val", + "foo==val", + "toggle=:", + "string=:", + "toggle=1 : foo", + "toggle=100", + "toggle==1", + "flags=+mu-lame : num=42: toggle=0", + "num=42 : string=blahblah", + "rational=0 : rational=1/2 : rational=1/-1", + "rational=-1/0", + "size=1024x768", + "size=pal", + "size=bogus", + "pix_fmt=yuv420p", + "pix_fmt=2", + "pix_fmt=bogus", + "sample_fmt=s16", + "sample_fmt=2", + "sample_fmt=bogus", + "video_rate=pal", + "video_rate=25", + "video_rate=30000/1001", + "video_rate=30/1.001", + "video_rate=bogus", + "duration=bogus", + "duration=123.45", + "duration=1\\:23\\:45.67", + "color=blue", + "color=0x223300", + "color=0x42FF07AA", + "cl=stereo+downmix", + "cl=foo", + "bin=boguss", + "bin=111", + "bin=ffff", + "num64=bogus", + "num64=44", + "num64=44.4", + "num64=-1", + "num64=101", + "flt=bogus", + "flt=2", + "flt=2.2", + "flt=-1", + "flt=101", + "dbl=bogus", + "dbl=2", + "dbl=2.2", + "dbl=-1", + "dbl=101", + "bool1=true", + "bool2=auto", + }; - av_opt_set_defaults(&test_ctx); - test_ctx.string = av_strdup("default"); - if (!test_ctx.string) - return AVERROR(ENOMEM); + test_ctx.class = &test_class; + av_opt_set_defaults(&test_ctx); - av_log_set_level(AV_LOG_DEBUG); + av_log_set_level(AV_LOG_QUIET); + + for (i=0; i < FF_ARRAY_ELEMS(options); i++) { + int silence_log = !strcmp(options[i], "rational=-1/0"); // inf formating differs between platforms + av_log(&test_ctx, AV_LOG_DEBUG, "Setting options string '%s'\n", options[i]); + if (silence_log) + av_log_set_callback(NULL); + if (av_set_options_string(&test_ctx, options[i], "=", ":") < 0) + printf("Error '%s'\n", options[i]); + else + printf("OK '%s'\n", options[i]); + av_log_set_callback(log_callback_help); + } + av_opt_free(&test_ctx); + } + + printf("\nTesting av_opt_set_from_string()\n"); + { + TestContext test_ctx = { 0 }; + static const char * const options[] = { + "", + "5", + "5:hello", + "5:hello:size=pal", + "5:size=pal:hello", + ":", + "=", + " 5 : hello : size = pal ", + "a_very_long_option_name_that_will_need_to_be_ellipsized_around_here=42" + }; + static const char * const shorthand[] = { "num", "string", NULL }; + + test_ctx.class = &test_class; + av_opt_set_defaults(&test_ctx); + + av_log_set_level(AV_LOG_QUIET); - for (i = 0; i < FF_ARRAY_ELEMS(options); i++) { - av_log(&test_ctx, AV_LOG_DEBUG, "Setting options string '%s'\n", options[i]); - if (av_set_options_string(&test_ctx, options[i], "=", ":") < 0) - av_log(&test_ctx, AV_LOG_ERROR, "Error setting options string: '%s'\n", options[i]); - printf("\n"); + for (i=0; i < FF_ARRAY_ELEMS(options); i++) { + av_log(&test_ctx, AV_LOG_DEBUG, "Setting options string '%s'\n", options[i]); + if (av_opt_set_from_string(&test_ctx, options[i], shorthand, "=", ":") < 0) + printf("Error '%s'\n", options[i]); + else + printf("OK '%s'\n", options[i]); + } + av_opt_free(&test_ctx); } return 0; diff --git a/libavutil/opt.c b/libavutil/opt.c index 7cb3d66557..7e9ad8f4b2 100644 --- a/libavutil/opt.c +++ b/libavutil/opt.c @@ -2,20 +2,20 @@ * AVOptions * Copyright (c) 2005 Michael Niedermayer <michaelni@gmx.at> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -25,19 +25,30 @@ * @author Michael Niedermayer <michaelni@gmx.at> */ -#include "avstring.h" #include "avutil.h" +#include "avassert.h" +#include "avstring.h" +#include "channel_layout.h" #include "common.h" #include "dict.h" #include "eval.h" #include "log.h" +#include "parseutils.h" +#include "pixdesc.h" #include "mathematics.h" #include "opt.h" +#include "samplefmt.h" +#include "bprint.h" + +#include <float.h> const AVOption *av_opt_next(const void *obj, const AVOption *last) { - AVClass *class = *(AVClass **)obj; - if (!last && class->option && class->option[0].name) + const AVClass *class; + if (!obj) + return NULL; + class = *(const AVClass**)obj; + if (!last && class && class->option && class->option[0].name) return class->option; if (last && last[1].name) return ++last; @@ -48,11 +59,20 @@ static int read_number(const AVOption *o, const void *dst, double *num, int *den { switch (o->type) { case AV_OPT_TYPE_FLAGS: - *intnum = *(unsigned int *)dst; + *intnum = *(unsigned int*)dst; + return 0; + case AV_OPT_TYPE_PIXEL_FMT: + *intnum = *(enum AVPixelFormat *)dst; return 0; + case AV_OPT_TYPE_SAMPLE_FMT: + *intnum = *(enum AVSampleFormat *)dst; + return 0; + case AV_OPT_TYPE_BOOL: case AV_OPT_TYPE_INT: *intnum = *(int *)dst; return 0; + case AV_OPT_TYPE_CHANNEL_LAYOUT: + case AV_OPT_TYPE_DURATION: case AV_OPT_TYPE_INT64: *intnum = *(int64_t *)dst; return 0; @@ -66,6 +86,9 @@ static int read_number(const AVOption *o, const void *dst, double *num, int *den *intnum = ((AVRational *)dst)->num; *den = ((AVRational *)dst)->den; return 0; + case AV_OPT_TYPE_CONST: + *num = o->default_val.dbl; + return 0; } return AVERROR(EINVAL); } @@ -74,16 +97,35 @@ static int write_number(void *obj, const AVOption *o, void *dst, double num, int { if (o->type != AV_OPT_TYPE_FLAGS && (o->max * den < num * intnum || o->min * den > num * intnum)) { - av_log(obj, AV_LOG_ERROR, "Value %f for parameter '%s' out of range\n", - num * intnum / den, o->name); + num = den ? num * intnum / den : (num * intnum ? INFINITY : NAN); + av_log(obj, AV_LOG_ERROR, "Value %f for parameter '%s' out of range [%g - %g]\n", + num, o->name, o->min, o->max); return AVERROR(ERANGE); } + if (o->type == AV_OPT_TYPE_FLAGS) { + double d = num*intnum/den; + if (d < -1.5 || d > 0xFFFFFFFF+0.5 || (llrint(d*256) & 255)) { + av_log(obj, AV_LOG_ERROR, + "Value %f for parameter '%s' is not a valid set of 32bit integer flags\n", + num*intnum/den, o->name); + return AVERROR(ERANGE); + } + } switch (o->type) { + case AV_OPT_TYPE_PIXEL_FMT: + *(enum AVPixelFormat *)dst = llrint(num / den) * intnum; + break; + case AV_OPT_TYPE_SAMPLE_FMT: + *(enum AVSampleFormat *)dst = llrint(num / den) * intnum; + break; + case AV_OPT_TYPE_BOOL: case AV_OPT_TYPE_FLAGS: case AV_OPT_TYPE_INT: *(int *)dst = llrint(num / den) * intnum; break; + case AV_OPT_TYPE_DURATION: + case AV_OPT_TYPE_CHANNEL_LAYOUT: case AV_OPT_TYPE_INT64: *(int64_t *)dst = llrint(num / den) * intnum; break; @@ -91,9 +133,10 @@ static int write_number(void *obj, const AVOption *o, void *dst, double num, int *(float *)dst = num * intnum / den; break; case AV_OPT_TYPE_DOUBLE: - *(double *)dst = num * intnum / den; + *(double *)dst = num * intnum / den; break; case AV_OPT_TYPE_RATIONAL: + case AV_OPT_TYPE_VIDEO_RATE: if ((int) num == num) *(AVRational *)dst = (AVRational) { num *intnum, den }; else @@ -105,22 +148,7 @@ static int write_number(void *obj, const AVOption *o, void *dst, double num, int return 0; } -static const double const_values[] = { - M_PI, - M_E, - FF_QP2LAMBDA, - 0 -}; - -static const char *const const_names[] = { - "PI", - "E", - "QP2LAMBDA", - 0 -}; - -static int hexchar2int(char c) -{ +static int hexchar2int(char c) { if (c >= '0' && c <= '9') return c - '0'; if (c >= 'a' && c <= 'f') @@ -134,11 +162,14 @@ static int set_string_binary(void *obj, const AVOption *o, const char *val, uint { int *lendst = (int *)(dst + 1); uint8_t *bin, *ptr; - int len = strlen(val); + int len; av_freep(dst); *lendst = 0; + if (!val || !(len = strlen(val))) + return 0; + if (len & 1) return AVERROR(EINVAL); len /= 2; @@ -177,42 +208,69 @@ static int set_string(void *obj, const AVOption *o, const char *val, uint8_t **d static int set_string_number(void *obj, void *target_obj, const AVOption *o, const char *val, void *dst) { - int ret = 0, notfirst = 0; + int ret = 0; + int num, den; + char c; + + if (sscanf(val, "%d%*1[:/]%d%c", &num, &den, &c) == 2) { + if ((ret = write_number(obj, o, dst, 1, den, num)) >= 0) + return ret; + ret = 0; + } + for (;;) { - int i, den = 1; + int i = 0; char buf[256]; int cmd = 0; - double d, num = 1; + double d; int64_t intnum = 1; - i = 0; - if (*val == '+' || *val == '-') { - if (o->type == AV_OPT_TYPE_FLAGS) + if (o->type == AV_OPT_TYPE_FLAGS) { + if (*val == '+' || *val == '-') cmd = *(val++); - else if (!notfirst) - buf[i++] = *val; + for (; i < sizeof(buf) - 1 && val[i] && val[i] != '+' && val[i] != '-'; i++) + buf[i] = val[i]; + buf[i] = 0; } - for (; i < sizeof(buf) - 1 && val[i] && val[i] != '+' && val[i] != '-'; i++) - buf[i] = val[i]; - buf[i] = 0; - { - const AVOption *o_named = av_opt_find(target_obj, buf, o->unit, 0, 0); + const AVOption *o_named = av_opt_find(target_obj, i ? buf : val, o->unit, 0, 0); + int res; + int ci = 0; + double const_values[64]; + const char * const_names[64]; if (o_named && o_named->type == AV_OPT_TYPE_CONST) d = DEFAULT_NUMVAL(o_named); - else if (!strcmp(buf, "default")) - d = DEFAULT_NUMVAL(o); - else if (!strcmp(buf, "max")) - d = o->max; - else if (!strcmp(buf, "min")) - d = o->min; - else if (!strcmp(buf, "none")) - d = 0; - else if (!strcmp(buf, "all")) - d = ~0; else { - int res = av_expr_parse_and_eval(&d, buf, const_names, const_values, NULL, NULL, NULL, NULL, NULL, 0, obj); + if (o->unit) { + for (o_named = NULL; o_named = av_opt_next(target_obj, o_named); ) { + if (o_named->type == AV_OPT_TYPE_CONST && + o_named->unit && + !strcmp(o_named->unit, o->unit)) { + if (ci + 6 >= FF_ARRAY_ELEMS(const_values)) { + av_log(obj, AV_LOG_ERROR, "const_values array too small for %s\n", o->unit); + return AVERROR_PATCHWELCOME; + } + const_names [ci ] = o_named->name; + const_values[ci++] = DEFAULT_NUMVAL(o_named); + } + } + } + const_names [ci ] = "default"; + const_values[ci++] = DEFAULT_NUMVAL(o); + const_names [ci ] = "max"; + const_values[ci++] = o->max; + const_names [ci ] = "min"; + const_values[ci++] = o->min; + const_names [ci ] = "none"; + const_values[ci++] = 0; + const_names [ci ] = "all"; + const_values[ci++] = ~0; + const_names [ci] = NULL; + const_values[ci] = 0; + + res = av_expr_parse_and_eval(&d, i ? buf : val, const_names, + const_values, NULL, NULL, NULL, NULL, NULL, 0, obj); if (res < 0) { av_log(obj, AV_LOG_ERROR, "Unable to parse option value \"%s\"\n", val); return res; @@ -224,37 +282,173 @@ static int set_string_number(void *obj, void *target_obj, const AVOption *o, con if (cmd == '+') d = intnum | (int64_t)d; else if (cmd == '-') - d = intnum & ~(int64_t)d; - } else { - read_number(o, dst, &num, &den, &intnum); - if (cmd == '+') - d = notfirst * num * intnum / den + d; - else if (cmd == '-') - d = notfirst * num * intnum / den - d; + d = intnum &~(int64_t)d; } if ((ret = write_number(obj, o, dst, d, 1, 1)) < 0) return ret; val += i; - if (!*val) + if (!i || !*val) return 0; - notfirst = 1; } return 0; } +static int set_string_image_size(void *obj, const AVOption *o, const char *val, int *dst) +{ + int ret; + + if (!val || !strcmp(val, "none")) { + dst[0] = + dst[1] = 0; + return 0; + } + ret = av_parse_video_size(dst, dst + 1, val); + if (ret < 0) + av_log(obj, AV_LOG_ERROR, "Unable to parse option value \"%s\" as image size\n", val); + return ret; +} + +static int set_string_video_rate(void *obj, const AVOption *o, const char *val, AVRational *dst) +{ + int ret; + if (!val) { + ret = AVERROR(EINVAL); + } else { + ret = av_parse_video_rate(dst, val); + } + if (ret < 0) + av_log(obj, AV_LOG_ERROR, "Unable to parse option value \"%s\" as video rate\n", val); + return ret; +} + +static int set_string_color(void *obj, const AVOption *o, const char *val, uint8_t *dst) +{ + int ret; + + if (!val) { + return 0; + } else { + ret = av_parse_color(dst, val, -1, obj); + if (ret < 0) + av_log(obj, AV_LOG_ERROR, "Unable to parse option value \"%s\" as color\n", val); + return ret; + } + return 0; +} + +static const char *get_bool_name(int val) +{ + if (val < 0) + return "auto"; + return val ? "true" : "false"; +} + +static int set_string_bool(void *obj, const AVOption *o, const char *val, int *dst) +{ + int n; + + if (!val) + return 0; + + if (!strcmp(val, "auto")) { + n = -1; + } else if (av_match_name(val, "true,y,yes,enable,enabled,on")) { + n = 1; + } else if (av_match_name(val, "false,n,no,disable,disabled,off")) { + n = 0; + } else { + char *end = NULL; + n = strtol(val, &end, 10); + if (val + strlen(val) != end) + goto fail; + } + + if (n < o->min || n > o->max) + goto fail; + + *dst = n; + return 0; + +fail: + av_log(obj, AV_LOG_ERROR, "Unable to parse option value \"%s\" as boolean\n", val); + return AVERROR(EINVAL); +} + +static int set_string_fmt(void *obj, const AVOption *o, const char *val, uint8_t *dst, + int fmt_nb, int ((*get_fmt)(const char *)), const char *desc) +{ + int fmt, min, max; + + if (!val || !strcmp(val, "none")) { + fmt = -1; + } else { + fmt = get_fmt(val); + if (fmt == -1) { + char *tail; + fmt = strtol(val, &tail, 0); + if (*tail || (unsigned)fmt >= fmt_nb) { + av_log(obj, AV_LOG_ERROR, + "Unable to parse option value \"%s\" as %s\n", val, desc); + return AVERROR(EINVAL); + } + } + } + + min = FFMAX(o->min, -1); + max = FFMIN(o->max, fmt_nb-1); + + // hack for compatibility with old ffmpeg + if(min == 0 && max == 0) { + min = -1; + max = fmt_nb-1; + } + + if (fmt < min || fmt > max) { + av_log(obj, AV_LOG_ERROR, + "Value %d for parameter '%s' out of %s format range [%d - %d]\n", + fmt, o->name, desc, min, max); + return AVERROR(ERANGE); + } + + *(int *)dst = fmt; + return 0; +} + +static int set_string_pixel_fmt(void *obj, const AVOption *o, const char *val, uint8_t *dst) +{ + return set_string_fmt(obj, o, val, dst, + AV_PIX_FMT_NB, av_get_pix_fmt, "pixel format"); +} + +static int set_string_sample_fmt(void *obj, const AVOption *o, const char *val, uint8_t *dst) +{ + return set_string_fmt(obj, o, val, dst, + AV_SAMPLE_FMT_NB, av_get_sample_fmt, "sample format"); +} + int av_opt_set(void *obj, const char *name, const char *val, int search_flags) { + int ret = 0; void *dst, *target_obj; const AVOption *o = av_opt_find2(obj, name, NULL, 0, search_flags, &target_obj); if (!o || !target_obj) return AVERROR_OPTION_NOT_FOUND; - if (!val || o->flags & AV_OPT_FLAG_READONLY) + if (!val && (o->type != AV_OPT_TYPE_STRING && + o->type != AV_OPT_TYPE_PIXEL_FMT && o->type != AV_OPT_TYPE_SAMPLE_FMT && + o->type != AV_OPT_TYPE_IMAGE_SIZE && o->type != AV_OPT_TYPE_VIDEO_RATE && + o->type != AV_OPT_TYPE_DURATION && o->type != AV_OPT_TYPE_COLOR && + o->type != AV_OPT_TYPE_CHANNEL_LAYOUT && o->type != AV_OPT_TYPE_BOOL)) + return AVERROR(EINVAL); + + if (o->flags & AV_OPT_FLAG_READONLY) return AVERROR(EINVAL); dst = ((uint8_t *)target_obj) + o->offset; switch (o->type) { + case AV_OPT_TYPE_BOOL: + return set_string_bool(obj, o, val, dst); case AV_OPT_TYPE_STRING: return set_string(obj, o, val, dst); case AV_OPT_TYPE_BINARY: @@ -266,6 +460,39 @@ int av_opt_set(void *obj, const char *name, const char *val, int search_flags) case AV_OPT_TYPE_DOUBLE: case AV_OPT_TYPE_RATIONAL: return set_string_number(obj, target_obj, o, val, dst); + case AV_OPT_TYPE_IMAGE_SIZE: + return set_string_image_size(obj, o, val, dst); + case AV_OPT_TYPE_VIDEO_RATE: + return set_string_video_rate(obj, o, val, dst); + case AV_OPT_TYPE_PIXEL_FMT: + return set_string_pixel_fmt(obj, o, val, dst); + case AV_OPT_TYPE_SAMPLE_FMT: + return set_string_sample_fmt(obj, o, val, dst); + case AV_OPT_TYPE_DURATION: + if (!val) { + *(int64_t *)dst = 0; + return 0; + } else { + if ((ret = av_parse_time(dst, val, 1)) < 0) + av_log(obj, AV_LOG_ERROR, "Unable to parse option value \"%s\" as duration\n", val); + return ret; + } + break; + case AV_OPT_TYPE_COLOR: + return set_string_color(obj, o, val, dst); + case AV_OPT_TYPE_CHANNEL_LAYOUT: + if (!val || !strcmp(val, "none")) { + *(int64_t *)dst = 0; + } else { + int64_t cl = av_get_channel_layout(val); + if (!cl) { + av_log(obj, AV_LOG_ERROR, "Unable to parse option value \"%s\" as channel layout\n", val); + ret = AVERROR(EINVAL); + } + *(int64_t *)dst = cl; + return ret; + } + break; } av_log(obj, AV_LOG_ERROR, "Invalid option type.\n"); @@ -333,8 +560,8 @@ int av_opt_set_bin(void *obj, const char *name, const uint8_t *val, int len, int if (o->type != AV_OPT_TYPE_BINARY || o->flags & AV_OPT_FLAG_READONLY) return AVERROR(EINVAL); - ptr = av_malloc(len); - if (!ptr) + ptr = len ? av_malloc(len) : NULL; + if (len && !ptr) return AVERROR(ENOMEM); dst = (uint8_t **)(((uint8_t *)target_obj) + o->offset); @@ -343,11 +570,106 @@ int av_opt_set_bin(void *obj, const char *name, const uint8_t *val, int len, int av_free(*dst); *dst = ptr; *lendst = len; - memcpy(ptr, val, len); + if (len) + memcpy(ptr, val, len); return 0; } +int av_opt_set_image_size(void *obj, const char *name, int w, int h, int search_flags) +{ + void *target_obj; + const AVOption *o = av_opt_find2(obj, name, NULL, 0, search_flags, &target_obj); + + if (!o || !target_obj) + return AVERROR_OPTION_NOT_FOUND; + if (o->type != AV_OPT_TYPE_IMAGE_SIZE) { + av_log(obj, AV_LOG_ERROR, + "The value set by option '%s' is not an image size.\n", o->name); + return AVERROR(EINVAL); + } + if (w<0 || h<0) { + av_log(obj, AV_LOG_ERROR, + "Invalid negative size value %dx%d for size '%s'\n", w, h, o->name); + return AVERROR(EINVAL); + } + *(int *)(((uint8_t *)target_obj) + o->offset) = w; + *(int *)(((uint8_t *)target_obj+sizeof(int)) + o->offset) = h; + return 0; +} + +int av_opt_set_video_rate(void *obj, const char *name, AVRational val, int search_flags) +{ + void *target_obj; + const AVOption *o = av_opt_find2(obj, name, NULL, 0, search_flags, &target_obj); + + if (!o || !target_obj) + return AVERROR_OPTION_NOT_FOUND; + if (o->type != AV_OPT_TYPE_VIDEO_RATE) { + av_log(obj, AV_LOG_ERROR, + "The value set by option '%s' is not a video rate.\n", o->name); + return AVERROR(EINVAL); + } + if (val.num <= 0 || val.den <= 0) + return AVERROR(EINVAL); + return set_number(obj, name, val.num, val.den, 1, search_flags); +} + +static int set_format(void *obj, const char *name, int fmt, int search_flags, + enum AVOptionType type, const char *desc, int nb_fmts) +{ + void *target_obj; + const AVOption *o = av_opt_find2(obj, name, NULL, 0, + search_flags, &target_obj); + int min, max; + + if (!o || !target_obj) + return AVERROR_OPTION_NOT_FOUND; + if (o->type != type) { + av_log(obj, AV_LOG_ERROR, + "The value set by option '%s' is not a %s format", name, desc); + return AVERROR(EINVAL); + } + + min = FFMAX(o->min, -1); + max = FFMIN(o->max, nb_fmts-1); + + if (fmt < min || fmt > max) { + av_log(obj, AV_LOG_ERROR, + "Value %d for parameter '%s' out of %s format range [%d - %d]\n", + fmt, name, desc, min, max); + return AVERROR(ERANGE); + } + *(int *)(((uint8_t *)target_obj) + o->offset) = fmt; + return 0; +} + +int av_opt_set_pixel_fmt(void *obj, const char *name, enum AVPixelFormat fmt, int search_flags) +{ + return set_format(obj, name, fmt, search_flags, AV_OPT_TYPE_PIXEL_FMT, "pixel", AV_PIX_FMT_NB); +} + +int av_opt_set_sample_fmt(void *obj, const char *name, enum AVSampleFormat fmt, int search_flags) +{ + return set_format(obj, name, fmt, search_flags, AV_OPT_TYPE_SAMPLE_FMT, "sample", AV_SAMPLE_FMT_NB); +} + +int av_opt_set_channel_layout(void *obj, const char *name, int64_t cl, int search_flags) +{ + void *target_obj; + const AVOption *o = av_opt_find2(obj, name, NULL, 0, search_flags, &target_obj); + + if (!o || !target_obj) + return AVERROR_OPTION_NOT_FOUND; + if (o->type != AV_OPT_TYPE_CHANNEL_LAYOUT) { + av_log(obj, AV_LOG_ERROR, + "The value set by option '%s' is not a channel layout.\n", o->name); + return AVERROR(EINVAL); + } + *(int64_t *)(((uint8_t *)target_obj) + o->offset) = cl; + return 0; +} + int av_opt_set_dict_val(void *obj, const char *name, const AVDictionary *val, int search_flags) { @@ -367,20 +689,59 @@ int av_opt_set_dict_val(void *obj, const char *name, const AVDictionary *val, return 0; } +static void format_duration(char *buf, size_t size, int64_t d) +{ + char *e; + + av_assert0(size >= 25); + if (d < 0 && d != INT64_MIN) { + *(buf++) = '-'; + size--; + d = -d; + } + if (d == INT64_MAX) + snprintf(buf, size, "INT64_MAX"); + else if (d == INT64_MIN) + snprintf(buf, size, "INT64_MIN"); + else if (d > (int64_t)3600*1000000) + snprintf(buf, size, "%"PRId64":%02d:%02d.%06d", d / 3600000000, + (int)((d / 60000000) % 60), + (int)((d / 1000000) % 60), + (int)(d % 1000000)); + else if (d > 60*1000000) + snprintf(buf, size, "%d:%02d.%06d", + (int)(d / 60000000), + (int)((d / 1000000) % 60), + (int)(d % 1000000)); + else + snprintf(buf, size, "%d.%06d", + (int)(d / 1000000), + (int)(d % 1000000)); + e = buf + strlen(buf); + while (e > buf && e[-1] == '0') + *(--e) = 0; + if (e > buf && e[-1] == '.') + *(--e) = 0; +} + int av_opt_get(void *obj, const char *name, int search_flags, uint8_t **out_val) { void *dst, *target_obj; const AVOption *o = av_opt_find2(obj, name, NULL, 0, search_flags, &target_obj); uint8_t *bin, buf[128]; int len, i, ret; + int64_t i64; - if (!o || !target_obj) + if (!o || !target_obj || (o->offset<=0 && o->type != AV_OPT_TYPE_CONST)) return AVERROR_OPTION_NOT_FOUND; dst = (uint8_t *)target_obj + o->offset; buf[0] = 0; switch (o->type) { + case AV_OPT_TYPE_BOOL: + ret = snprintf(buf, sizeof(buf), "%s", (char *)av_x_if_null(get_bool_name(*(int *)dst), "invalid")); + break; case AV_OPT_TYPE_FLAGS: ret = snprintf(buf, sizeof(buf), "0x%08X", *(int *)dst); break; @@ -388,7 +749,7 @@ int av_opt_get(void *obj, const char *name, int search_flags, uint8_t **out_val) ret = snprintf(buf, sizeof(buf), "%d", *(int *)dst); break; case AV_OPT_TYPE_INT64: - ret = snprintf(buf, sizeof(buf), "%" PRId64, *(int64_t *)dst); + ret = snprintf(buf, sizeof(buf), "%"PRId64, *(int64_t *)dst); break; case AV_OPT_TYPE_FLOAT: ret = snprintf(buf, sizeof(buf), "%f", *(float *)dst); @@ -396,26 +757,64 @@ int av_opt_get(void *obj, const char *name, int search_flags, uint8_t **out_val) case AV_OPT_TYPE_DOUBLE: ret = snprintf(buf, sizeof(buf), "%f", *(double *)dst); break; + case AV_OPT_TYPE_VIDEO_RATE: case AV_OPT_TYPE_RATIONAL: - ret = snprintf(buf, sizeof(buf), "%d/%d", ((AVRational *)dst)->num, - ((AVRational *)dst)->den); + ret = snprintf(buf, sizeof(buf), "%d/%d", ((AVRational *)dst)->num, ((AVRational *)dst)->den); + break; + case AV_OPT_TYPE_CONST: + ret = snprintf(buf, sizeof(buf), "%f", o->default_val.dbl); break; case AV_OPT_TYPE_STRING: - if (*(uint8_t **)dst) + if (*(uint8_t **)dst) { *out_val = av_strdup(*(uint8_t **)dst); - else + } else if (search_flags & AV_OPT_ALLOW_NULL) { + *out_val = NULL; + return 0; + } else { *out_val = av_strdup(""); + } return *out_val ? 0 : AVERROR(ENOMEM); case AV_OPT_TYPE_BINARY: + if (!*(uint8_t **)dst && (search_flags & AV_OPT_ALLOW_NULL)) { + *out_val = NULL; + return 0; + } len = *(int *)(((uint8_t *)dst) + sizeof(uint8_t *)); if ((uint64_t)len * 2 + 1 > INT_MAX) return AVERROR(EINVAL); if (!(*out_val = av_malloc(len * 2 + 1))) return AVERROR(ENOMEM); + if (!len) { + *out_val[0] = '\0'; + return 0; + } bin = *(uint8_t **)dst; for (i = 0; i < len; i++) snprintf(*out_val + i * 2, 3, "%02X", bin[i]); return 0; + case AV_OPT_TYPE_IMAGE_SIZE: + ret = snprintf(buf, sizeof(buf), "%dx%d", ((int *)dst)[0], ((int *)dst)[1]); + break; + case AV_OPT_TYPE_PIXEL_FMT: + ret = snprintf(buf, sizeof(buf), "%s", (char *)av_x_if_null(av_get_pix_fmt_name(*(enum AVPixelFormat *)dst), "none")); + break; + case AV_OPT_TYPE_SAMPLE_FMT: + ret = snprintf(buf, sizeof(buf), "%s", (char *)av_x_if_null(av_get_sample_fmt_name(*(enum AVSampleFormat *)dst), "none")); + break; + case AV_OPT_TYPE_DURATION: + i64 = *(int64_t *)dst; + format_duration(buf, sizeof(buf), i64); + ret = strlen(buf); // no overflow possible, checked by an assert + break; + case AV_OPT_TYPE_COLOR: + ret = snprintf(buf, sizeof(buf), "0x%02x%02x%02x%02x", + (int)((uint8_t *)dst)[0], (int)((uint8_t *)dst)[1], + (int)((uint8_t *)dst)[2], (int)((uint8_t *)dst)[3]); + break; + case AV_OPT_TYPE_CHANNEL_LAYOUT: + i64 = *(int64_t *)dst; + ret = snprintf(buf, sizeof(buf), "0x%"PRIx64, i64); + break; default: return AVERROR(EINVAL); } @@ -426,7 +825,7 @@ int av_opt_get(void *obj, const char *name, int search_flags, uint8_t **out_val) return *out_val ? 0 : AVERROR(ENOMEM); } -static int get_number(void *obj, const char *name, double *num, int *den, int64_t *intnum, +static int get_number(void *obj, const char *name, const AVOption **o_out, double *num, int *den, int64_t *intnum, int search_flags) { void *dst, *target_obj; @@ -436,6 +835,8 @@ static int get_number(void *obj, const char *name, double *num, int *den, int64_ dst = ((uint8_t *)target_obj) + o->offset; + if (o_out) *o_out= o; + return read_number(o, dst, num, den, intnum); error: @@ -450,7 +851,7 @@ int av_opt_get_int(void *obj, const char *name, int search_flags, int64_t *out_v double num = 1; int ret, den = 1; - if ((ret = get_number(obj, name, &num, &den, &intnum, search_flags)) < 0) + if ((ret = get_number(obj, name, NULL, &num, &den, &intnum, search_flags)) < 0) return ret; *out_val = num * intnum / den; return 0; @@ -462,7 +863,7 @@ int av_opt_get_double(void *obj, const char *name, int search_flags, double *out double num = 1; int ret, den = 1; - if ((ret = get_number(obj, name, &num, &den, &intnum, search_flags)) < 0) + if ((ret = get_number(obj, name, NULL, &num, &den, &intnum, search_flags)) < 0) return ret; *out_val = num * intnum / den; return 0; @@ -474,7 +875,41 @@ int av_opt_get_q(void *obj, const char *name, int search_flags, AVRational *out_ double num = 1; int ret, den = 1; - if ((ret = get_number(obj, name, &num, &den, &intnum, search_flags)) < 0) + if ((ret = get_number(obj, name, NULL, &num, &den, &intnum, search_flags)) < 0) + return ret; + + if (num == 1.0 && (int)intnum == intnum) + *out_val = (AVRational){intnum, den}; + else + *out_val = av_d2q(num*intnum/den, 1<<24); + return 0; +} + +int av_opt_get_image_size(void *obj, const char *name, int search_flags, int *w_out, int *h_out) +{ + void *dst, *target_obj; + const AVOption *o = av_opt_find2(obj, name, NULL, 0, search_flags, &target_obj); + if (!o || !target_obj) + return AVERROR_OPTION_NOT_FOUND; + if (o->type != AV_OPT_TYPE_IMAGE_SIZE) { + av_log(obj, AV_LOG_ERROR, + "The value for option '%s' is not an image size.\n", name); + return AVERROR(EINVAL); + } + + dst = ((uint8_t*)target_obj) + o->offset; + if (w_out) *w_out = *(int *)dst; + if (h_out) *h_out = *((int *)dst+1); + return 0; +} + +int av_opt_get_video_rate(void *obj, const char *name, int search_flags, AVRational *out_val) +{ + int64_t intnum = 1; + double num = 1; + int ret, den = 1; + + if ((ret = get_number(obj, name, NULL, &num, &den, &intnum, search_flags)) < 0) return ret; if (num == 1.0 && (int)intnum == intnum) @@ -484,6 +919,51 @@ int av_opt_get_q(void *obj, const char *name, int search_flags, AVRational *out_ return 0; } +static int get_format(void *obj, const char *name, int search_flags, int *out_fmt, + enum AVOptionType type, const char *desc) +{ + void *dst, *target_obj; + const AVOption *o = av_opt_find2(obj, name, NULL, 0, search_flags, &target_obj); + if (!o || !target_obj) + return AVERROR_OPTION_NOT_FOUND; + if (o->type != type) { + av_log(obj, AV_LOG_ERROR, + "The value for option '%s' is not a %s format.\n", desc, name); + return AVERROR(EINVAL); + } + + dst = ((uint8_t*)target_obj) + o->offset; + *out_fmt = *(int *)dst; + return 0; +} + +int av_opt_get_pixel_fmt(void *obj, const char *name, int search_flags, enum AVPixelFormat *out_fmt) +{ + return get_format(obj, name, search_flags, out_fmt, AV_OPT_TYPE_PIXEL_FMT, "pixel"); +} + +int av_opt_get_sample_fmt(void *obj, const char *name, int search_flags, enum AVSampleFormat *out_fmt) +{ + return get_format(obj, name, search_flags, out_fmt, AV_OPT_TYPE_SAMPLE_FMT, "sample"); +} + +int av_opt_get_channel_layout(void *obj, const char *name, int search_flags, int64_t *cl) +{ + void *dst, *target_obj; + const AVOption *o = av_opt_find2(obj, name, NULL, 0, search_flags, &target_obj); + if (!o || !target_obj) + return AVERROR_OPTION_NOT_FOUND; + if (o->type != AV_OPT_TYPE_CHANNEL_LAYOUT) { + av_log(obj, AV_LOG_ERROR, + "The value for option '%s' is not a channel layout.\n", name); + return AVERROR(EINVAL); + } + + dst = ((uint8_t*)target_obj) + o->offset; + *cl = *(int64_t *)dst; + return 0; +} + int av_opt_get_dict_val(void *obj, const char *name, int search_flags, AVDictionary **out_val) { void *target_obj; @@ -514,10 +994,79 @@ int av_opt_flag_is_set(void *obj, const char *field_name, const char *flag_name) return res & flag->default_val.i64; } +static void log_value(void *av_log_obj, int level, double d) +{ + if (d == INT_MAX) { + av_log(av_log_obj, level, "INT_MAX"); + } else if (d == INT_MIN) { + av_log(av_log_obj, level, "INT_MIN"); + } else if (d == UINT32_MAX) { + av_log(av_log_obj, level, "UINT32_MAX"); + } else if (d == (double)INT64_MAX) { + av_log(av_log_obj, level, "I64_MAX"); + } else if (d == INT64_MIN) { + av_log(av_log_obj, level, "I64_MIN"); + } else if (d == FLT_MAX) { + av_log(av_log_obj, level, "FLT_MAX"); + } else if (d == FLT_MIN) { + av_log(av_log_obj, level, "FLT_MIN"); + } else if (d == -FLT_MAX) { + av_log(av_log_obj, level, "-FLT_MAX"); + } else if (d == -FLT_MIN) { + av_log(av_log_obj, level, "-FLT_MIN"); + } else if (d == DBL_MAX) { + av_log(av_log_obj, level, "DBL_MAX"); + } else if (d == DBL_MIN) { + av_log(av_log_obj, level, "DBL_MIN"); + } else if (d == -DBL_MAX) { + av_log(av_log_obj, level, "-DBL_MAX"); + } else if (d == -DBL_MIN) { + av_log(av_log_obj, level, "-DBL_MIN"); + } else { + av_log(av_log_obj, level, "%g", d); + } +} + +static const char *get_opt_const_name(void *obj, const char *unit, int64_t value) +{ + const AVOption *opt = NULL; + + if (!unit) + return NULL; + while ((opt = av_opt_next(obj, opt))) + if (opt->type == AV_OPT_TYPE_CONST && !strcmp(opt->unit, unit) && + opt->default_val.i64 == value) + return opt->name; + return NULL; +} + +static char *get_opt_flags_string(void *obj, const char *unit, int64_t value) +{ + const AVOption *opt = NULL; + char flags[512]; + + flags[0] = 0; + if (!unit) + return NULL; + while ((opt = av_opt_next(obj, opt))) { + if (opt->type == AV_OPT_TYPE_CONST && !strcmp(opt->unit, unit) && + opt->default_val.i64 & value) { + if (flags[0]) + av_strlcatf(flags, sizeof(flags), "+"); + av_strlcatf(flags, sizeof(flags), "%s", opt->name); + } + } + if (flags[0]) + return av_strdup(flags); + return NULL; +} + static void opt_list(void *obj, void *av_log_obj, const char *unit, int req_flags, int rej_flags) { const AVOption *opt = NULL; + AVOptionRanges *r; + int i; while ((opt = av_opt_next(obj, opt))) { if (!(opt->flags & req_flags) || (opt->flags & rej_flags)) @@ -534,50 +1083,161 @@ static void opt_list(void *obj, void *av_log_obj, const char *unit, else if (unit && opt->type == AV_OPT_TYPE_CONST && strcmp(unit, opt->unit)) continue; else if (unit && opt->type == AV_OPT_TYPE_CONST) - av_log(av_log_obj, AV_LOG_INFO, " %-15s ", opt->name); + av_log(av_log_obj, AV_LOG_INFO, " %-15s ", opt->name); else - av_log(av_log_obj, AV_LOG_INFO, "-%-17s ", opt->name); + av_log(av_log_obj, AV_LOG_INFO, " %s%-17s ", + (opt->flags & AV_OPT_FLAG_FILTERING_PARAM) ? "" : "-", + opt->name); switch (opt->type) { - case AV_OPT_TYPE_FLAGS: - av_log(av_log_obj, AV_LOG_INFO, "%-7s ", "<flags>"); - break; - case AV_OPT_TYPE_INT: - av_log(av_log_obj, AV_LOG_INFO, "%-7s ", "<int>"); - break; - case AV_OPT_TYPE_INT64: - av_log(av_log_obj, AV_LOG_INFO, "%-7s ", "<int64>"); - break; - case AV_OPT_TYPE_DOUBLE: - av_log(av_log_obj, AV_LOG_INFO, "%-7s ", "<double>"); - break; - case AV_OPT_TYPE_FLOAT: - av_log(av_log_obj, AV_LOG_INFO, "%-7s ", "<float>"); - break; - case AV_OPT_TYPE_STRING: - av_log(av_log_obj, AV_LOG_INFO, "%-7s ", "<string>"); - break; - case AV_OPT_TYPE_RATIONAL: - av_log(av_log_obj, AV_LOG_INFO, "%-7s ", "<rational>"); - break; - case AV_OPT_TYPE_BINARY: - av_log(av_log_obj, AV_LOG_INFO, "%-7s ", "<binary>"); - break; - case AV_OPT_TYPE_CONST: - default: - av_log(av_log_obj, AV_LOG_INFO, "%-7s ", ""); - break; + case AV_OPT_TYPE_FLAGS: + av_log(av_log_obj, AV_LOG_INFO, "%-12s ", "<flags>"); + break; + case AV_OPT_TYPE_INT: + av_log(av_log_obj, AV_LOG_INFO, "%-12s ", "<int>"); + break; + case AV_OPT_TYPE_INT64: + av_log(av_log_obj, AV_LOG_INFO, "%-12s ", "<int64>"); + break; + case AV_OPT_TYPE_DOUBLE: + av_log(av_log_obj, AV_LOG_INFO, "%-12s ", "<double>"); + break; + case AV_OPT_TYPE_FLOAT: + av_log(av_log_obj, AV_LOG_INFO, "%-12s ", "<float>"); + break; + case AV_OPT_TYPE_STRING: + av_log(av_log_obj, AV_LOG_INFO, "%-12s ", "<string>"); + break; + case AV_OPT_TYPE_RATIONAL: + av_log(av_log_obj, AV_LOG_INFO, "%-12s ", "<rational>"); + break; + case AV_OPT_TYPE_BINARY: + av_log(av_log_obj, AV_LOG_INFO, "%-12s ", "<binary>"); + break; + case AV_OPT_TYPE_IMAGE_SIZE: + av_log(av_log_obj, AV_LOG_INFO, "%-12s ", "<image_size>"); + break; + case AV_OPT_TYPE_VIDEO_RATE: + av_log(av_log_obj, AV_LOG_INFO, "%-12s ", "<video_rate>"); + break; + case AV_OPT_TYPE_PIXEL_FMT: + av_log(av_log_obj, AV_LOG_INFO, "%-12s ", "<pix_fmt>"); + break; + case AV_OPT_TYPE_SAMPLE_FMT: + av_log(av_log_obj, AV_LOG_INFO, "%-12s ", "<sample_fmt>"); + break; + case AV_OPT_TYPE_DURATION: + av_log(av_log_obj, AV_LOG_INFO, "%-12s ", "<duration>"); + break; + case AV_OPT_TYPE_COLOR: + av_log(av_log_obj, AV_LOG_INFO, "%-12s ", "<color>"); + break; + case AV_OPT_TYPE_CHANNEL_LAYOUT: + av_log(av_log_obj, AV_LOG_INFO, "%-12s ", "<channel_layout>"); + break; + case AV_OPT_TYPE_BOOL: + av_log(av_log_obj, AV_LOG_INFO, "%-12s ", "<boolean>"); + break; + case AV_OPT_TYPE_CONST: + default: + av_log(av_log_obj, AV_LOG_INFO, "%-12s ", ""); + break; } av_log(av_log_obj, AV_LOG_INFO, "%c", (opt->flags & AV_OPT_FLAG_ENCODING_PARAM) ? 'E' : '.'); av_log(av_log_obj, AV_LOG_INFO, "%c", (opt->flags & AV_OPT_FLAG_DECODING_PARAM) ? 'D' : '.'); - av_log(av_log_obj, AV_LOG_INFO, "%c", (opt->flags & AV_OPT_FLAG_VIDEO_PARAM) ? 'V' : '.'); - av_log(av_log_obj, AV_LOG_INFO, "%c", (opt->flags & AV_OPT_FLAG_AUDIO_PARAM) ? 'A' : '.'); + av_log(av_log_obj, AV_LOG_INFO, "%c", (opt->flags & AV_OPT_FLAG_FILTERING_PARAM)? 'F' : '.'); + av_log(av_log_obj, AV_LOG_INFO, "%c", (opt->flags & AV_OPT_FLAG_VIDEO_PARAM ) ? 'V' : '.'); + av_log(av_log_obj, AV_LOG_INFO, "%c", (opt->flags & AV_OPT_FLAG_AUDIO_PARAM ) ? 'A' : '.'); av_log(av_log_obj, AV_LOG_INFO, "%c", (opt->flags & AV_OPT_FLAG_SUBTITLE_PARAM) ? 'S' : '.'); av_log(av_log_obj, AV_LOG_INFO, "%c", (opt->flags & AV_OPT_FLAG_EXPORT) ? 'X' : '.'); av_log(av_log_obj, AV_LOG_INFO, "%c", (opt->flags & AV_OPT_FLAG_READONLY) ? 'R' : '.'); if (opt->help) av_log(av_log_obj, AV_LOG_INFO, " %s", opt->help); + + if (av_opt_query_ranges(&r, obj, opt->name, AV_OPT_SEARCH_FAKE_OBJ) >= 0) { + switch (opt->type) { + case AV_OPT_TYPE_INT: + case AV_OPT_TYPE_INT64: + case AV_OPT_TYPE_DOUBLE: + case AV_OPT_TYPE_FLOAT: + case AV_OPT_TYPE_RATIONAL: + for (i = 0; i < r->nb_ranges; i++) { + av_log(av_log_obj, AV_LOG_INFO, " (from "); + log_value(av_log_obj, AV_LOG_INFO, r->range[i]->value_min); + av_log(av_log_obj, AV_LOG_INFO, " to "); + log_value(av_log_obj, AV_LOG_INFO, r->range[i]->value_max); + av_log(av_log_obj, AV_LOG_INFO, ")"); + } + break; + } + av_opt_freep_ranges(&r); + } + + if (opt->type != AV_OPT_TYPE_CONST && + opt->type != AV_OPT_TYPE_BINARY && + !((opt->type == AV_OPT_TYPE_COLOR || + opt->type == AV_OPT_TYPE_IMAGE_SIZE || + opt->type == AV_OPT_TYPE_STRING || + opt->type == AV_OPT_TYPE_VIDEO_RATE) && + !opt->default_val.str)) { + av_log(av_log_obj, AV_LOG_INFO, " (default "); + switch (opt->type) { + case AV_OPT_TYPE_BOOL: + av_log(av_log_obj, AV_LOG_INFO, "%s", (char *)av_x_if_null(get_bool_name(opt->default_val.i64), "invalid")); + break; + case AV_OPT_TYPE_FLAGS: { + char *def_flags = get_opt_flags_string(obj, opt->unit, opt->default_val.i64); + if (def_flags) { + av_log(av_log_obj, AV_LOG_INFO, "%s", def_flags); + av_freep(&def_flags); + } else { + av_log(av_log_obj, AV_LOG_INFO, "%"PRIX64, opt->default_val.i64); + } + break; + } + case AV_OPT_TYPE_DURATION: { + char buf[25]; + format_duration(buf, sizeof(buf), opt->default_val.i64); + av_log(av_log_obj, AV_LOG_INFO, "%s", buf); + break; + } + case AV_OPT_TYPE_INT: + case AV_OPT_TYPE_INT64: { + const char *def_const = get_opt_const_name(obj, opt->unit, opt->default_val.i64); + if (def_const) + av_log(av_log_obj, AV_LOG_INFO, "%s", def_const); + else + log_value(av_log_obj, AV_LOG_INFO, opt->default_val.i64); + break; + } + case AV_OPT_TYPE_DOUBLE: + case AV_OPT_TYPE_FLOAT: + log_value(av_log_obj, AV_LOG_INFO, opt->default_val.dbl); + break; + case AV_OPT_TYPE_RATIONAL: { + AVRational q = av_d2q(opt->default_val.dbl, INT_MAX); + av_log(av_log_obj, AV_LOG_INFO, "%d/%d", q.num, q.den); } + break; + case AV_OPT_TYPE_PIXEL_FMT: + av_log(av_log_obj, AV_LOG_INFO, "%s", (char *)av_x_if_null(av_get_pix_fmt_name(opt->default_val.i64), "none")); + break; + case AV_OPT_TYPE_SAMPLE_FMT: + av_log(av_log_obj, AV_LOG_INFO, "%s", (char *)av_x_if_null(av_get_sample_fmt_name(opt->default_val.i64), "none")); + break; + case AV_OPT_TYPE_COLOR: + case AV_OPT_TYPE_IMAGE_SIZE: + case AV_OPT_TYPE_STRING: + case AV_OPT_TYPE_VIDEO_RATE: + av_log(av_log_obj, AV_LOG_INFO, "\"%s\"", opt->default_val.str); + break; + case AV_OPT_TYPE_CHANNEL_LAYOUT: + av_log(av_log_obj, AV_LOG_INFO, "0x%"PRIx64, opt->default_val.i64); + break; + } + av_log(av_log_obj, AV_LOG_INFO, ")"); + } + av_log(av_log_obj, AV_LOG_INFO, "\n"); if (opt->unit && opt->type != AV_OPT_TYPE_CONST) opt_list(obj, av_log_obj, opt->unit, req_flags, rej_flags); @@ -598,41 +1258,65 @@ int av_opt_show2(void *obj, void *av_log_obj, int req_flags, int rej_flags) void av_opt_set_defaults(void *s) { + av_opt_set_defaults2(s, 0, 0); +} + +void av_opt_set_defaults2(void *s, int mask, int flags) +{ const AVOption *opt = NULL; while ((opt = av_opt_next(s, opt))) { + void *dst = ((uint8_t*)s) + opt->offset; + + if ((opt->flags & mask) != flags) + continue; + if (opt->flags & AV_OPT_FLAG_READONLY) continue; switch (opt->type) { - case AV_OPT_TYPE_CONST: - /* Nothing to be done here */ - break; - case AV_OPT_TYPE_FLAGS: - case AV_OPT_TYPE_INT: - case AV_OPT_TYPE_INT64: - av_opt_set_int(s, opt->name, opt->default_val.i64, 0); + case AV_OPT_TYPE_CONST: + /* Nothing to be done here */ + break; + case AV_OPT_TYPE_BOOL: + case AV_OPT_TYPE_FLAGS: + case AV_OPT_TYPE_INT: + case AV_OPT_TYPE_INT64: + case AV_OPT_TYPE_DURATION: + case AV_OPT_TYPE_CHANNEL_LAYOUT: + case AV_OPT_TYPE_PIXEL_FMT: + case AV_OPT_TYPE_SAMPLE_FMT: + write_number(s, opt, dst, 1, 1, opt->default_val.i64); + break; + case AV_OPT_TYPE_DOUBLE: + case AV_OPT_TYPE_FLOAT: { + double val; + val = opt->default_val.dbl; + write_number(s, opt, dst, val, 1, 1); + } break; - case AV_OPT_TYPE_DOUBLE: - case AV_OPT_TYPE_FLOAT: - { - double val; - val = opt->default_val.dbl; - av_opt_set_double(s, opt->name, val, 0); - } - break; - case AV_OPT_TYPE_RATIONAL: - { - AVRational val; - val = av_d2q(opt->default_val.dbl, INT_MAX); - av_opt_set_q(s, opt->name, val, 0); - } - break; - case AV_OPT_TYPE_STRING: - av_opt_set(s, opt->name, opt->default_val.str, 0); + case AV_OPT_TYPE_RATIONAL: { + AVRational val; + val = av_d2q(opt->default_val.dbl, INT_MAX); + write_number(s, opt, dst, 1, val.den, val.num); + } break; - case AV_OPT_TYPE_BINARY: - case AV_OPT_TYPE_DICT: - /* Cannot set defaults for these types */ + case AV_OPT_TYPE_COLOR: + set_string_color(s, opt, opt->default_val.str, dst); + break; + case AV_OPT_TYPE_STRING: + set_string(s, opt, opt->default_val.str, dst); + break; + case AV_OPT_TYPE_IMAGE_SIZE: + set_string_image_size(s, opt, opt->default_val.str, dst); + break; + case AV_OPT_TYPE_VIDEO_RATE: + set_string_video_rate(s, opt, opt->default_val.str, dst); + break; + case AV_OPT_TYPE_BINARY: + set_string_binary(s, opt, opt->default_val.str, dst); + break; + case AV_OPT_TYPE_DICT: + /* Cannot set defaults for these types */ break; default: av_log(s, AV_LOG_DEBUG, "AVOption type %d of option %s not implemented yet\n", @@ -681,7 +1365,7 @@ static int parse_key_value_pair(void *ctx, const char **buf, return AVERROR(EINVAL); } - av_log(ctx, AV_LOG_DEBUG, "Setting value '%s' for key '%s'\n", val, key); + av_log(ctx, AV_LOG_DEBUG, "Setting entry with key '%s' to value '%s'\n", key, val); ret = av_opt_set(ctx, key, val, AV_OPT_SEARCH_CHILDREN); if (ret == AVERROR_OPTION_NOT_FOUND) @@ -712,6 +1396,118 @@ int av_set_options_string(void *ctx, const char *opts, return count; } +#define WHITESPACES " \n\t\r" + +static int is_key_char(char c) +{ + return (unsigned)((c | 32) - 'a') < 26 || + (unsigned)(c - '0') < 10 || + c == '-' || c == '_' || c == '/' || c == '.'; +} + +/** + * Read a key from a string. + * + * The key consists of is_key_char characters and must be terminated by a + * character from the delim string; spaces are ignored. + * + * @return 0 for success (even with ellipsis), <0 for failure + */ +static int get_key(const char **ropts, const char *delim, char **rkey) +{ + const char *opts = *ropts; + const char *key_start, *key_end; + + key_start = opts += strspn(opts, WHITESPACES); + while (is_key_char(*opts)) + opts++; + key_end = opts; + opts += strspn(opts, WHITESPACES); + if (!*opts || !strchr(delim, *opts)) + return AVERROR(EINVAL); + opts++; + if (!(*rkey = av_malloc(key_end - key_start + 1))) + return AVERROR(ENOMEM); + memcpy(*rkey, key_start, key_end - key_start); + (*rkey)[key_end - key_start] = 0; + *ropts = opts; + return 0; +} + +int av_opt_get_key_value(const char **ropts, + const char *key_val_sep, const char *pairs_sep, + unsigned flags, + char **rkey, char **rval) +{ + int ret; + char *key = NULL, *val; + const char *opts = *ropts; + + if ((ret = get_key(&opts, key_val_sep, &key)) < 0 && + !(flags & AV_OPT_FLAG_IMPLICIT_KEY)) + return AVERROR(EINVAL); + if (!(val = av_get_token(&opts, pairs_sep))) { + av_free(key); + return AVERROR(ENOMEM); + } + *ropts = opts; + *rkey = key; + *rval = val; + return 0; +} + +int av_opt_set_from_string(void *ctx, const char *opts, + const char *const *shorthand, + const char *key_val_sep, const char *pairs_sep) +{ + int ret, count = 0; + const char *dummy_shorthand = NULL; + char *av_uninit(parsed_key), *av_uninit(value); + const char *key; + + if (!opts) + return 0; + if (!shorthand) + shorthand = &dummy_shorthand; + + while (*opts) { + ret = av_opt_get_key_value(&opts, key_val_sep, pairs_sep, + *shorthand ? AV_OPT_FLAG_IMPLICIT_KEY : 0, + &parsed_key, &value); + if (ret < 0) { + if (ret == AVERROR(EINVAL)) + av_log(ctx, AV_LOG_ERROR, "No option name near '%s'\n", opts); + else + av_log(ctx, AV_LOG_ERROR, "Unable to parse '%s': %s\n", opts, + av_err2str(ret)); + return ret; + } + if (*opts) + opts++; + if (parsed_key) { + key = parsed_key; + while (*shorthand) /* discard all remaining shorthand */ + shorthand++; + } else { + key = *(shorthand++); + } + + av_log(ctx, AV_LOG_DEBUG, "Setting '%s' to value '%s'\n", key, value); + if ((ret = av_opt_set(ctx, key, value, 0)) < 0) { + if (ret == AVERROR_OPTION_NOT_FOUND) + av_log(ctx, AV_LOG_ERROR, "Option '%s' not found\n", key); + av_free(value); + av_free(parsed_key); + return ret; + } + + av_free(value); + av_free(parsed_key); + count++; + } + return count; +} + void av_opt_free(void *obj) { const AVOption *o = NULL; @@ -732,19 +1528,23 @@ void av_opt_free(void *obj) } } -int av_opt_set_dict(void *obj, AVDictionary **options) +int av_opt_set_dict2(void *obj, AVDictionary **options, int search_flags) { AVDictionaryEntry *t = NULL; AVDictionary *tmp = NULL; int ret = 0; + if (!options) + return 0; + while ((t = av_dict_get(*options, "", t, AV_DICT_IGNORE_SUFFIX))) { - ret = av_opt_set(obj, t->key, t->value, 0); + ret = av_opt_set(obj, t->key, t->value, search_flags); if (ret == AVERROR_OPTION_NOT_FOUND) - av_dict_set(&tmp, t->key, t->value, 0); - else if (ret < 0) { + ret = av_dict_set(&tmp, t->key, t->value, 0); + if (ret < 0) { av_log(obj, AV_LOG_ERROR, "Error setting option %s to value %s.\n", t->key, t->value); - break; + av_dict_free(&tmp); + return ret; } ret = 0; } @@ -753,6 +1553,11 @@ int av_opt_set_dict(void *obj, AVDictionary **options) return ret; } +int av_opt_set_dict(void *obj, AVDictionary **options) +{ + return av_opt_set_dict2(obj, options, 0); +} + const AVOption *av_opt_find(void *obj, const char *name, const char *unit, int opt_flags, int search_flags) { @@ -762,9 +1567,14 @@ const AVOption *av_opt_find(void *obj, const char *name, const char *unit, const AVOption *av_opt_find2(void *obj, const char *name, const char *unit, int opt_flags, int search_flags, void **target_obj) { - const AVClass *c = *(AVClass **)obj; + const AVClass *c; const AVOption *o = NULL; + if(!obj) + return NULL; + + c= *(AVClass**)obj; + if (!c) return NULL; @@ -785,7 +1595,7 @@ const AVOption *av_opt_find2(void *obj, const char *name, const char *unit, while (o = av_opt_next(obj, o)) { if (!strcmp(o->name, name) && (o->flags & opt_flags) == opt_flags && ((!unit && o->type != AV_OPT_TYPE_CONST) || - (unit && o->unit && !strcmp(o->unit, unit)))) { + (unit && o->type == AV_OPT_TYPE_CONST && o->unit && !strcmp(o->unit, unit)))) { if (target_obj) { if (!(search_flags & AV_OPT_SEARCH_FAKE_OBJ)) *target_obj = obj; @@ -813,12 +1623,23 @@ const AVClass *av_opt_child_class_next(const AVClass *parent, const AVClass *pre return NULL; } +void *av_opt_ptr(const AVClass *class, void *obj, const char *name) +{ + const AVOption *opt= av_opt_find2(&class, name, NULL, 0, AV_OPT_SEARCH_FAKE_OBJ, NULL); + if(!opt) + return NULL; + return (uint8_t*)obj + opt->offset; +} + static int opt_size(enum AVOptionType type) { - switch (type) { + switch(type) { + case AV_OPT_TYPE_BOOL: case AV_OPT_TYPE_INT: case AV_OPT_TYPE_FLAGS: return sizeof(int); + case AV_OPT_TYPE_DURATION: + case AV_OPT_TYPE_CHANNEL_LAYOUT: case AV_OPT_TYPE_INT64: return sizeof(int64_t); case AV_OPT_TYPE_DOUBLE: @@ -826,11 +1647,20 @@ static int opt_size(enum AVOptionType type) case AV_OPT_TYPE_FLOAT: return sizeof(float); case AV_OPT_TYPE_STRING: - return sizeof(uint8_t *); + return sizeof(uint8_t*); + case AV_OPT_TYPE_VIDEO_RATE: case AV_OPT_TYPE_RATIONAL: return sizeof(AVRational); case AV_OPT_TYPE_BINARY: - return sizeof(uint8_t *) + sizeof(int); + return sizeof(uint8_t*) + sizeof(int); + case AV_OPT_TYPE_IMAGE_SIZE: + return sizeof(int[2]); + case AV_OPT_TYPE_PIXEL_FMT: + return sizeof(enum AVPixelFormat); + case AV_OPT_TYPE_SAMPLE_FMT: + return sizeof(enum AVSampleFormat); + case AV_OPT_TYPE_COLOR: + return 4; } return AVERROR(EINVAL); } @@ -855,26 +1685,32 @@ int av_opt_copy(void *dst, const void *src) uint8_t **field_src8 = (uint8_t **)field_src; if (o->type == AV_OPT_TYPE_STRING) { - set_string(dst, o, *field_src8, field_dst8); + if (*field_dst8 != *field_src8) + av_freep(field_dst8); + *field_dst8 = av_strdup(*field_src8); if (*field_src8 && !*field_dst8) ret = AVERROR(ENOMEM); } else if (o->type == AV_OPT_TYPE_BINARY) { int len = *(int *)(field_src8 + 1); if (*field_dst8 != *field_src8) av_freep(field_dst8); - if (len) { - *field_dst8 = av_malloc(len); - if (!*field_dst8) { - ret = AVERROR(ENOMEM); - len = 0; - } - memcpy(*field_dst8, *field_src8, len); - } else { - *field_dst8 = NULL; + *field_dst8 = av_memdup(*field_src8, len); + if (len && !*field_dst8) { + ret = AVERROR(ENOMEM); + len = 0; } *(int *)(field_dst8 + 1) = len; } else if (o->type == AV_OPT_TYPE_CONST) { // do nothing + } else if (o->type == AV_OPT_TYPE_DICT) { + AVDictionary **sdict = (AVDictionary **) field_src; + AVDictionary **ddict = (AVDictionary **) field_dst; + if (*sdict != *ddict) + av_dict_free(ddict); + *ddict = NULL; + av_dict_copy(ddict, *sdict, 0); + if (av_dict_count(*sdict) != av_dict_count(*ddict)) + ret = AVERROR(ENOMEM); } else { int size = opt_size(o->type); if (size < 0) @@ -885,3 +1721,269 @@ int av_opt_copy(void *dst, const void *src) } return ret; } + +int av_opt_query_ranges(AVOptionRanges **ranges_arg, void *obj, const char *key, int flags) +{ + int ret; + const AVClass *c = *(AVClass**)obj; + int (*callback)(AVOptionRanges **, void *obj, const char *key, int flags) = NULL; + + if (c->version > (52 << 16 | 11 << 8)) + callback = c->query_ranges; + + if (!callback) + callback = av_opt_query_ranges_default; + + ret = callback(ranges_arg, obj, key, flags); + if (ret >= 0) { + if (!(flags & AV_OPT_MULTI_COMPONENT_RANGE)) + ret = 1; + (*ranges_arg)->nb_components = ret; + } + return ret; +} + +int av_opt_query_ranges_default(AVOptionRanges **ranges_arg, void *obj, const char *key, int flags) +{ + AVOptionRanges *ranges = av_mallocz(sizeof(*ranges)); + AVOptionRange **range_array = av_mallocz(sizeof(void*)); + AVOptionRange *range = av_mallocz(sizeof(*range)); + const AVOption *field = av_opt_find(obj, key, NULL, 0, flags); + int ret; + + *ranges_arg = NULL; + + if (!ranges || !range || !range_array || !field) { + ret = AVERROR(ENOMEM); + goto fail; + } + + ranges->range = range_array; + ranges->range[0] = range; + ranges->nb_ranges = 1; + ranges->nb_components = 1; + range->is_range = 1; + range->value_min = field->min; + range->value_max = field->max; + + switch (field->type) { + case AV_OPT_TYPE_BOOL: + case AV_OPT_TYPE_INT: + case AV_OPT_TYPE_INT64: + case AV_OPT_TYPE_PIXEL_FMT: + case AV_OPT_TYPE_SAMPLE_FMT: + case AV_OPT_TYPE_FLOAT: + case AV_OPT_TYPE_DOUBLE: + case AV_OPT_TYPE_DURATION: + case AV_OPT_TYPE_COLOR: + case AV_OPT_TYPE_CHANNEL_LAYOUT: + break; + case AV_OPT_TYPE_STRING: + range->component_min = 0; + range->component_max = 0x10FFFF; // max unicode value + range->value_min = -1; + range->value_max = INT_MAX; + break; + case AV_OPT_TYPE_RATIONAL: + range->component_min = INT_MIN; + range->component_max = INT_MAX; + break; + case AV_OPT_TYPE_IMAGE_SIZE: + range->component_min = 0; + range->component_max = INT_MAX/128/8; + range->value_min = 0; + range->value_max = INT_MAX/8; + break; + case AV_OPT_TYPE_VIDEO_RATE: + range->component_min = 1; + range->component_max = INT_MAX; + range->value_min = 1; + range->value_max = INT_MAX; + break; + default: + ret = AVERROR(ENOSYS); + goto fail; + } + + *ranges_arg = ranges; + return 1; +fail: + av_free(ranges); + av_free(range); + av_free(range_array); + return ret; +} + +void av_opt_freep_ranges(AVOptionRanges **rangesp) +{ + int i; + AVOptionRanges *ranges = *rangesp; + + if (!ranges) + return; + + for (i = 0; i < ranges->nb_ranges * ranges->nb_components; i++) { + AVOptionRange *range = ranges->range[i]; + if (range) { + av_freep(&range->str); + av_freep(&ranges->range[i]); + } + } + av_freep(&ranges->range); + av_freep(rangesp); +} + +int av_opt_is_set_to_default(void *obj, const AVOption *o) +{ + int64_t i64; + double d, d2; + float f; + AVRational q; + int ret, w, h; + char *str; + void *dst; + + if (!o || !obj) + return AVERROR(EINVAL); + + dst = ((uint8_t*)obj) + o->offset; + + switch (o->type) { + case AV_OPT_TYPE_CONST: + return 1; + case AV_OPT_TYPE_BOOL: + case AV_OPT_TYPE_FLAGS: + case AV_OPT_TYPE_PIXEL_FMT: + case AV_OPT_TYPE_SAMPLE_FMT: + case AV_OPT_TYPE_INT: + case AV_OPT_TYPE_CHANNEL_LAYOUT: + case AV_OPT_TYPE_DURATION: + case AV_OPT_TYPE_INT64: + read_number(o, dst, NULL, NULL, &i64); + return o->default_val.i64 == i64; + case AV_OPT_TYPE_STRING: + str = *(char **)dst; + if (str == o->default_val.str) //2 NULLs + return 1; + if (!str || !o->default_val.str) //1 NULL + return 0; + return !strcmp(str, o->default_val.str); + case AV_OPT_TYPE_DOUBLE: + read_number(o, dst, &d, NULL, NULL); + return o->default_val.dbl == d; + case AV_OPT_TYPE_FLOAT: + read_number(o, dst, &d, NULL, NULL); + f = o->default_val.dbl; + d2 = f; + return d2 == d; + case AV_OPT_TYPE_RATIONAL: + q = av_d2q(o->default_val.dbl, INT_MAX); + return !av_cmp_q(*(AVRational*)dst, q); + case AV_OPT_TYPE_BINARY: { + struct { + uint8_t *data; + int size; + } tmp = {0}; + int opt_size = *(int *)((void **)dst + 1); + void *opt_ptr = *(void **)dst; + if (!opt_size && (!o->default_val.str || !strlen(o->default_val.str))) + return 1; + if (!opt_size || !o->default_val.str || !strlen(o->default_val.str )) + return 0; + if (opt_size != strlen(o->default_val.str) / 2) + return 0; + ret = set_string_binary(NULL, NULL, o->default_val.str, &tmp.data); + if (!ret) + ret = !memcmp(opt_ptr, tmp.data, tmp.size); + av_free(tmp.data); + return ret; + } + case AV_OPT_TYPE_DICT: + /* Binary and dict have not default support yet. Any pointer is not default. */ + return !!(*(void **)dst); + case AV_OPT_TYPE_IMAGE_SIZE: + if (!o->default_val.str || !strcmp(o->default_val.str, "none")) + w = h = 0; + else if ((ret = av_parse_video_size(&w, &h, o->default_val.str)) < 0) + return ret; + return (w == *(int *)dst) && (h == *((int *)dst+1)); + case AV_OPT_TYPE_VIDEO_RATE: + q = (AVRational){0, 0}; + if (o->default_val.str) { + if ((ret = av_parse_video_rate(&q, o->default_val.str)) < 0) + return ret; + } + return !av_cmp_q(*(AVRational*)dst, q); + case AV_OPT_TYPE_COLOR: { + uint8_t color[4] = {0, 0, 0, 0}; + if (o->default_val.str) { + if ((ret = av_parse_color(color, o->default_val.str, -1, NULL)) < 0) + return ret; + } + return !memcmp(color, dst, sizeof(color)); + } + default: + av_log(obj, AV_LOG_WARNING, "Not supported option type: %d, option name: %s\n", o->type, o->name); + break; + } + return AVERROR_PATCHWELCOME; +} + +int av_opt_is_set_to_default_by_name(void *obj, const char *name, int search_flags) +{ + const AVOption *o; + void *target; + if (!obj) + return AVERROR(EINVAL); + o = av_opt_find2(obj, name, NULL, 0, search_flags, &target); + if (!o) + return AVERROR_OPTION_NOT_FOUND; + return av_opt_is_set_to_default(target, o); +} + +int av_opt_serialize(void *obj, int opt_flags, int flags, char **buffer, + const char key_val_sep, const char pairs_sep) +{ + const AVOption *o = NULL; + uint8_t *buf; + AVBPrint bprint; + int ret, cnt = 0; + const char special_chars[] = {pairs_sep, key_val_sep, '\0'}; + + if (pairs_sep == '\0' || key_val_sep == '\0' || pairs_sep == key_val_sep || + pairs_sep == '\\' || key_val_sep == '\\') { + av_log(obj, AV_LOG_ERROR, "Invalid separator(s) found."); + return AVERROR(EINVAL); + } + + if (!obj || !buffer) + return AVERROR(EINVAL); + + *buffer = NULL; + av_bprint_init(&bprint, 64, AV_BPRINT_SIZE_UNLIMITED); + + while (o = av_opt_next(obj, o)) { + if (o->type == AV_OPT_TYPE_CONST) + continue; + if ((flags & AV_OPT_SERIALIZE_OPT_FLAGS_EXACT) && o->flags != opt_flags) + continue; + else if (((o->flags & opt_flags) != opt_flags)) + continue; + if (flags & AV_OPT_SERIALIZE_SKIP_DEFAULTS && av_opt_is_set_to_default(obj, o) > 0) + continue; + if ((ret = av_opt_get(obj, o->name, 0, &buf)) < 0) { + av_bprint_finalize(&bprint, NULL); + return ret; + } + if (buf) { + if (cnt++) + av_bprint_append_data(&bprint, &pairs_sep, 1); + av_bprint_escape(&bprint, o->name, special_chars, AV_ESCAPE_MODE_BACKSLASH, 0); + av_bprint_append_data(&bprint, &key_val_sep, 1); + av_bprint_escape(&bprint, buf, special_chars, AV_ESCAPE_MODE_BACKSLASH, 0); + av_freep(&buf); + } + } + av_bprint_finalize(&bprint, buffer); + return 0; +} diff --git a/libavutil/opt.h b/libavutil/opt.h index 906b869d51..753434d628 100644 --- a/libavutil/opt.h +++ b/libavutil/opt.h @@ -2,20 +2,20 @@ * AVOptions * copyright (c) 2005 Michael Niedermayer <michaelni@gmx.at> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -31,6 +31,9 @@ #include "avutil.h" #include "dict.h" #include "log.h" +#include "pixfmt.h" +#include "samplefmt.h" +#include "version.h" /** * @defgroup avoptions AVOptions @@ -44,7 +47,7 @@ * This section describes how to add AVOptions capabilities to a struct. * * All AVOptions-related information is stored in an AVClass. Therefore - * the first member of the struct must be a pointer to an AVClass describing it. + * the first member of the struct should be a pointer to an AVClass describing it. * The option field of the AVClass must be set to a NULL-terminated static array * of AVOptions. Each AVOption must have a non-empty name, a type, a default * value and for number-type AVOptions also a range of allowed values. It must @@ -81,7 +84,7 @@ * @endcode * * Next, when allocating your struct, you must ensure that the AVClass pointer - * is set to the correct value. Then, av_opt_set_defaults() must be called to + * is set to the correct value. Then, av_opt_set_defaults() can be called to * initialize defaults. After that the struct is ready to be used with the * AVOptions API. * @@ -161,7 +164,7 @@ * * @subsection avoptions_implement_named_constants Named constants * It is possible to create named constants for options. Simply set the unit - * field of the option the constants should apply to to a string and + * field of the option the constants should apply to a string and * create the constants themselves as options of type AV_OPT_TYPE_CONST * with their unit field set to the same string. * Their default_val field should contain the value of the named @@ -176,7 +179,7 @@ * * @section avoptions_use Using AVOptions * This section deals with accessing options in an AVOptions-enabled struct. - * Such structs in Libav are e.g. AVCodecContext in libavcodec or + * Such structs in FFmpeg are e.g. AVCodecContext in libavcodec or * AVFormatContext in libavformat. * * @subsection avoptions_use_examine Examining AVOptions @@ -210,7 +213,7 @@ * In some cases it may be more convenient to put all options into an * AVDictionary and call av_opt_set_dict() on it. A specific case of this * are the format/codec open functions in lavf/lavc which take a dictionary - * filled with option as a parameter. This allows to set some options + * filled with option as a parameter. This makes it possible to set some options * that cannot be set otherwise, since e.g. the input file format is not known * before the file is actually opened. */ @@ -226,6 +229,14 @@ enum AVOptionType{ AV_OPT_TYPE_BINARY, ///< offset must point to a pointer immediately followed by an int for the length AV_OPT_TYPE_DICT, AV_OPT_TYPE_CONST = 128, + AV_OPT_TYPE_IMAGE_SIZE = MKBETAG('S','I','Z','E'), ///< offset must point to two consecutive integers + AV_OPT_TYPE_PIXEL_FMT = MKBETAG('P','F','M','T'), + AV_OPT_TYPE_SAMPLE_FMT = MKBETAG('S','F','M','T'), + AV_OPT_TYPE_VIDEO_RATE = MKBETAG('V','R','A','T'), ///< offset must point to AVRational + AV_OPT_TYPE_DURATION = MKBETAG('D','U','R',' '), + AV_OPT_TYPE_COLOR = MKBETAG('C','O','L','R'), + AV_OPT_TYPE_CHANNEL_LAYOUT = MKBETAG('C','H','L','A'), + AV_OPT_TYPE_BOOL = MKBETAG('B','O','O','L'), }; /** @@ -278,6 +289,7 @@ typedef struct AVOption { * This flag only makes sense when AV_OPT_FLAG_EXPORT is also set. */ #define AV_OPT_FLAG_READONLY 128 +#define AV_OPT_FLAG_FILTERING_PARAM (1<<16) ///< a generic parameter which can be set by the user for filtering //FIXME think about enc-audio, ... style flags /** @@ -289,6 +301,74 @@ typedef struct AVOption { } AVOption; /** + * A single allowed range of values, or a single allowed value. + */ +typedef struct AVOptionRange { + const char *str; + /** + * Value range. + * For string ranges this represents the min/max length. + * For dimensions this represents the min/max pixel count or width/height in multi-component case. + */ + double value_min, value_max; + /** + * Value's component range. + * For string this represents the unicode range for chars, 0-127 limits to ASCII. + */ + double component_min, component_max; + /** + * Range flag. + * If set to 1 the struct encodes a range, if set to 0 a single value. + */ + int is_range; +} AVOptionRange; + +/** + * List of AVOptionRange structs. + */ +typedef struct AVOptionRanges { + /** + * Array of option ranges. + * + * Most of option types use just one component. + * Following describes multi-component option types: + * + * AV_OPT_TYPE_IMAGE_SIZE: + * component index 0: range of pixel count (width * height). + * component index 1: range of width. + * component index 2: range of height. + * + * @note To obtain multi-component version of this structure, user must + * provide AV_OPT_MULTI_COMPONENT_RANGE to av_opt_query_ranges or + * av_opt_query_ranges_default function. + * + * Multi-component range can be read as in following example: + * + * @code + * int range_index, component_index; + * AVOptionRanges *ranges; + * AVOptionRange *range[3]; //may require more than 3 in the future. + * av_opt_query_ranges(&ranges, obj, key, AV_OPT_MULTI_COMPONENT_RANGE); + * for (range_index = 0; range_index < ranges->nb_ranges; range_index++) { + * for (component_index = 0; component_index < ranges->nb_components; component_index++) + * range[component_index] = ranges->range[ranges->nb_ranges * component_index + range_index]; + * //do something with range here. + * } + * av_opt_freep_ranges(&ranges); + * @endcode + */ + AVOptionRange **range; + /** + * Number of ranges per component. + */ + int nb_ranges; + /** + * Number of componentes. + */ + int nb_components; +} AVOptionRanges; + +/** * Show the obj options. * * @param req_flags requested flags for the options to show. Show only the @@ -307,11 +387,23 @@ int av_opt_show2(void *obj, void *av_log_obj, int req_flags, int rej_flags); void av_opt_set_defaults(void *s); /** + * Set the values of all AVOption fields to their default values. Only these + * AVOption fields for which (opt->flags & mask) == flags will have their + * default applied to s. + * + * @param s an AVOption-enabled struct (its first member must be a pointer to AVClass) + * @param mask combination of AV_OPT_FLAG_* + * @param flags combination of AV_OPT_FLAG_* + */ +void av_opt_set_defaults2(void *s, int mask, int flags); + +/** * Parse the key/value pairs list in opts. For each key/value pair * found, stores the value in the field in ctx that is named like the * key. ctx must be an AVClass context, storing is done using * AVOptions. * + * @param opts options string to parse, may be NULL * @param key_val_sep a 0-terminated list of characters used to * separate key from value * @param pairs_sep a 0-terminated list of characters used to separate @@ -326,6 +418,36 @@ int av_set_options_string(void *ctx, const char *opts, const char *key_val_sep, const char *pairs_sep); /** + * Parse the key-value pairs list in opts. For each key=value pair found, + * set the value of the corresponding option in ctx. + * + * @param ctx the AVClass object to set options on + * @param opts the options string, key-value pairs separated by a + * delimiter + * @param shorthand a NULL-terminated array of options names for shorthand + * notation: if the first field in opts has no key part, + * the key is taken from the first element of shorthand; + * then again for the second, etc., until either opts is + * finished, shorthand is finished or a named option is + * found; after that, all options must be named + * @param key_val_sep a 0-terminated list of characters used to separate + * key from value, for example '=' + * @param pairs_sep a 0-terminated list of characters used to separate + * two pairs from each other, for example ':' or ',' + * @return the number of successfully set key=value pairs, or a negative + * value corresponding to an AVERROR code in case of error: + * AVERROR(EINVAL) if opts cannot be parsed, + * the error code issued by av_set_string3() if a key/value pair + * cannot be set + * + * Options names must use only the following characters: a-z A-Z 0-9 - . / _ + * Separators must use characters distinct from option names and from each + * other. + */ +int av_opt_set_from_string(void *ctx, const char *opts, + const char *const *shorthand, + const char *key_val_sep, const char *pairs_sep); +/** * Free all allocated objects in obj. */ void av_opt_free(void *obj); @@ -340,7 +462,7 @@ void av_opt_free(void *obj); */ int av_opt_flag_is_set(void *obj, const char *field_name, const char *flag_name); -/* +/** * Set all the options from a given dictionary on an object. * * @param obj a struct whose first element is a pointer to AVClass @@ -356,6 +478,57 @@ int av_opt_flag_is_set(void *obj, const char *field_name, const char *flag_name) */ int av_opt_set_dict(void *obj, struct AVDictionary **options); + +/** + * Set all the options from a given dictionary on an object. + * + * @param obj a struct whose first element is a pointer to AVClass + * @param options options to process. This dictionary will be freed and replaced + * by a new one containing all options not found in obj. + * Of course this new dictionary needs to be freed by caller + * with av_dict_free(). + * @param search_flags A combination of AV_OPT_SEARCH_*. + * + * @return 0 on success, a negative AVERROR if some option was found in obj, + * but could not be set. + * + * @see av_dict_copy() + */ +int av_opt_set_dict2(void *obj, struct AVDictionary **options, int search_flags); + +/** + * Extract a key-value pair from the beginning of a string. + * + * @param ropts pointer to the options string, will be updated to + * point to the rest of the string (one of the pairs_sep + * or the final NUL) + * @param key_val_sep a 0-terminated list of characters used to separate + * key from value, for example '=' + * @param pairs_sep a 0-terminated list of characters used to separate + * two pairs from each other, for example ':' or ',' + * @param flags flags; see the AV_OPT_FLAG_* values below + * @param rkey parsed key; must be freed using av_free() + * @param rval parsed value; must be freed using av_free() + * + * @return >=0 for success, or a negative value corresponding to an + * AVERROR code in case of error; in particular: + * AVERROR(EINVAL) if no key is present + * + */ +int av_opt_get_key_value(const char **ropts, + const char *key_val_sep, const char *pairs_sep, + unsigned flags, + char **rkey, char **rval); + +enum { + + /** + * Accept to parse a value without a key; the key will then be returned + * as NULL. + */ + AV_OPT_FLAG_IMPLICIT_KEY = 1, +}; + /** * @defgroup opt_eval_funcs Evaluating option strings * @{ @@ -380,15 +553,28 @@ int av_opt_eval_q (void *obj, const AVOption *o, const char *val, AVRational * @} */ -#define AV_OPT_SEARCH_CHILDREN 0x0001 /**< Search in possible children of the - given object first. */ +#define AV_OPT_SEARCH_CHILDREN (1 << 0) /**< Search in possible children of the + given object first. */ /** * The obj passed to av_opt_find() is fake -- only a double pointer to AVClass * instead of a required pointer to a struct containing AVClass. This is * useful for searching for options without needing to allocate the corresponding * object. */ -#define AV_OPT_SEARCH_FAKE_OBJ 0x0002 +#define AV_OPT_SEARCH_FAKE_OBJ (1 << 1) + +/** + * In av_opt_get, return NULL if the option has a pointer type and is set to NULL, + * rather than returning an empty string. + */ +#define AV_OPT_ALLOW_NULL (1 << 2) + +/** + * Allows av_opt_query_ranges and av_opt_query_ranges_default to return more than + * one component for certain option types. + * @see AVOptionRanges for details. + */ +#define AV_OPT_MULTI_COMPONENT_RANGE (1 << 12) /** * Look for an option in an object. Consider only options which @@ -497,11 +683,33 @@ int av_opt_set_int (void *obj, const char *name, int64_t val, int search int av_opt_set_double (void *obj, const char *name, double val, int search_flags); int av_opt_set_q (void *obj, const char *name, AVRational val, int search_flags); int av_opt_set_bin (void *obj, const char *name, const uint8_t *val, int size, int search_flags); +int av_opt_set_image_size(void *obj, const char *name, int w, int h, int search_flags); +int av_opt_set_pixel_fmt (void *obj, const char *name, enum AVPixelFormat fmt, int search_flags); +int av_opt_set_sample_fmt(void *obj, const char *name, enum AVSampleFormat fmt, int search_flags); +int av_opt_set_video_rate(void *obj, const char *name, AVRational val, int search_flags); +int av_opt_set_channel_layout(void *obj, const char *name, int64_t ch_layout, int search_flags); /** * @note Any old dictionary present is discarded and replaced with a copy of the new one. The * caller still owns val is and responsible for freeing it. */ int av_opt_set_dict_val(void *obj, const char *name, const AVDictionary *val, int search_flags); + +/** + * Set a binary option to an integer list. + * + * @param obj AVClass object to set options on + * @param name name of the binary option + * @param val pointer to an integer list (must have the correct type with + * regard to the contents of the list) + * @param term list terminator (usually 0 or -1) + * @param flags search flags + */ +#define av_opt_set_int_list(obj, name, val, term, flags) \ + (av_int_list_length(val, term) > INT_MAX / sizeof(*(val)) ? \ + AVERROR(EINVAL) : \ + av_opt_set_bin(obj, name, (const uint8_t *)(val), \ + av_int_list_length(val, term) * sizeof(*(val)), flags)) + /** * @} */ @@ -516,20 +724,61 @@ int av_opt_set_dict_val(void *obj, const char *name, const AVDictionary *val, in * @param[in] search_flags flags passed to av_opt_find2. I.e. if AV_OPT_SEARCH_CHILDREN * is passed here, then the option may be found in a child of obj. * @param[out] out_val value of the option will be written here - * @return 0 on success, a negative error code otherwise + * @return >=0 on success, a negative error code otherwise */ /** - * @note the returned string will av_malloc()ed and must be av_free()ed by the caller + * @note the returned string will be av_malloc()ed and must be av_free()ed by the caller + * + * @note if AV_OPT_ALLOW_NULL is set in search_flags in av_opt_get, and the option has + * AV_OPT_TYPE_STRING or AV_OPT_TYPE_BINARY and is set to NULL, *out_val will be set + * to NULL instead of an allocated empty string. */ int av_opt_get (void *obj, const char *name, int search_flags, uint8_t **out_val); int av_opt_get_int (void *obj, const char *name, int search_flags, int64_t *out_val); int av_opt_get_double (void *obj, const char *name, int search_flags, double *out_val); int av_opt_get_q (void *obj, const char *name, int search_flags, AVRational *out_val); +int av_opt_get_image_size(void *obj, const char *name, int search_flags, int *w_out, int *h_out); +int av_opt_get_pixel_fmt (void *obj, const char *name, int search_flags, enum AVPixelFormat *out_fmt); +int av_opt_get_sample_fmt(void *obj, const char *name, int search_flags, enum AVSampleFormat *out_fmt); +int av_opt_get_video_rate(void *obj, const char *name, int search_flags, AVRational *out_val); +int av_opt_get_channel_layout(void *obj, const char *name, int search_flags, int64_t *ch_layout); /** * @param[out] out_val The returned dictionary is a copy of the actual value and must * be freed with av_dict_free() by the caller */ int av_opt_get_dict_val(void *obj, const char *name, int search_flags, AVDictionary **out_val); +/** + * @} + */ +/** + * Gets a pointer to the requested field in a struct. + * This function allows accessing a struct even when its fields are moved or + * renamed since the application making the access has been compiled, + * + * @returns a pointer to the field, it can be cast to the correct type and read + * or written to. + */ +void *av_opt_ptr(const AVClass *avclass, void *obj, const char *name); + +/** + * Free an AVOptionRanges struct and set it to NULL. + */ +void av_opt_freep_ranges(AVOptionRanges **ranges); + +/** + * Get a list of allowed ranges for the given option. + * + * The returned list may depend on other fields in obj like for example profile. + * + * @param flags is a bitmask of flags, undefined flags should not be set and should be ignored + * AV_OPT_SEARCH_FAKE_OBJ indicates that the obj is a double pointer to a AVClass instead of a full instance + * AV_OPT_MULTI_COMPONENT_RANGE indicates that function may return more than one component, @see AVOptionRanges + * + * The result must be freed with av_opt_freep_ranges. + * + * @return number of compontents returned on success, a negative errro code otherwise + */ +int av_opt_query_ranges(AVOptionRanges **, void *obj, const char *key, int flags); /** * Copy options from src object into dest object. @@ -544,7 +793,72 @@ int av_opt_get_dict_val(void *obj, const char *name, int search_flags, AVDiction int av_opt_copy(void *dest, const void *src); /** - * @} + * Get a default list of allowed ranges for the given option. + * + * This list is constructed without using the AVClass.query_ranges() callback + * and can be used as fallback from within the callback. + * + * @param flags is a bitmask of flags, undefined flags should not be set and should be ignored + * AV_OPT_SEARCH_FAKE_OBJ indicates that the obj is a double pointer to a AVClass instead of a full instance + * AV_OPT_MULTI_COMPONENT_RANGE indicates that function may return more than one component, @see AVOptionRanges + * + * The result must be freed with av_opt_free_ranges. + * + * @return number of compontents returned on success, a negative errro code otherwise + */ +int av_opt_query_ranges_default(AVOptionRanges **, void *obj, const char *key, int flags); + +/** + * Check if given option is set to its default value. + * + * Options o must belong to the obj. This function must not be called to check child's options state. + * @see av_opt_is_set_to_default_by_name(). + * + * @param obj AVClass object to check option on + * @param o option to be checked + * @return >0 when option is set to its default, + * 0 when option is not set its default, + * <0 on error + */ +int av_opt_is_set_to_default(void *obj, const AVOption *o); + +/** + * Check if given option is set to its default value. + * + * @param obj AVClass object to check option on + * @param name option name + * @param search_flags combination of AV_OPT_SEARCH_* + * @return >0 when option is set to its default, + * 0 when option is not set its default, + * <0 on error + */ +int av_opt_is_set_to_default_by_name(void *obj, const char *name, int search_flags); + + +#define AV_OPT_SERIALIZE_SKIP_DEFAULTS 0x00000001 ///< Serialize options that are not set to default values only. +#define AV_OPT_SERIALIZE_OPT_FLAGS_EXACT 0x00000002 ///< Serialize options that exactly match opt_flags only. + +/** + * Serialize object's options. + * + * Create a string containing object's serialized options. + * Such string may be passed back to av_opt_set_from_string() in order to restore option values. + * A key/value or pairs separator occurring in the serialized value or + * name string are escaped through the av_escape() function. + * + * @param[in] obj AVClass object to serialize + * @param[in] opt_flags serialize options with all the specified flags set (AV_OPT_FLAG) + * @param[in] flags combination of AV_OPT_SERIALIZE_* flags + * @param[out] buffer Pointer to buffer that will be allocated with string containg serialized options. + * Buffer must be freed by the caller when is no longer needed. + * @param[in] key_val_sep character used to separate key from value + * @param[in] pairs_sep character used to separate two pairs from each other + * @return >= 0 on success, negative on error + * @warning Separators cannot be neither '\\' nor '\0'. They also cannot be the same. + */ +int av_opt_serialize(void *obj, int opt_flags, int flags, char **buffer, + const char key_val_sep, const char pairs_sep); +/** * @} */ diff --git a/libavutil/parseutils-test.c b/libavutil/parseutils-test.c index 149861f1d5..fa9c1c1d05 100644 --- a/libavutil/parseutils-test.c +++ b/libavutil/parseutils-test.c @@ -1,33 +1,41 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ +#define TEST +#include "parseutils.c" + #include <stdint.h> #include <stdio.h> #include "common.h" #include "log.h" #include "rational.h" -#include "parseutils.h" -int main(void) +static uint32_t randomv = MKTAG('L','A','V','U'); + +static uint32_t av_get_random_seed_deterministic(void) +{ + return randomv = randomv * 1664525 + 1013904223; +} + +static void test_av_parse_video_rate(void) { int i; - uint8_t rgba[4]; static const char *const rates[] = { "-inf", "inf", @@ -55,7 +63,23 @@ int main(void) " 21332.2324 ", " -21332.2324 ", }; + + for (i = 0; i < FF_ARRAY_ELEMS(rates); i++) { + int ret; + AVRational q = { 0, 0 }; + ret = av_parse_video_rate(&q, rates[i]); + printf("'%s' -> %d/%d %s\n", + rates[i], q.num, q.den, ret ? "ERROR" : "OK"); + } +} + +static void test_av_parse_color(void) +{ + int i; + uint8_t rgba[4]; static const char *const color_names[] = { + "bikeshed", + "RaNdOm", "foo", "red", "Red ", @@ -92,25 +116,145 @@ int main(void) "red@-0.0", }; - printf("Testing av_parse_video_rate()\n"); + av_log_set_level(AV_LOG_DEBUG); - for (i = 0; i < FF_ARRAY_ELEMS(rates); i++) { - int ret; - AVRational q = { 0, 0 }; - ret = av_parse_video_rate(&q, rates[i]); - printf("'%s' -> %d/%d %s\n", - rates[i], q.num, q.den, ret ? "ERROR" : "OK"); + for (i = 0; i < FF_ARRAY_ELEMS(color_names); i++) { + if (av_parse_color(rgba, color_names[i], -1, NULL) >= 0) + printf("%s -> R(%d) G(%d) B(%d) A(%d)\n", + color_names[i], rgba[0], rgba[1], rgba[2], rgba[3]); + else + printf("%s -> error\n", color_names[i]); } +} - printf("\nTesting av_parse_color()\n"); +static void test_av_small_strptime(void) +{ + int i; + struct tm tm = { 0 }; + struct fmt_timespec_entry { + const char *fmt, *timespec; + } fmt_timespec_entries[] = { + { "%Y-%m-%d", "2012-12-21" }, + { "%Y - %m - %d", "2012-12-21" }, + { "%Y-%m-%d %H:%M:%S", "2012-12-21 20:12:21" }, + { " %Y - %m - %d %H : %M : %S", " 2012 - 12 - 21 20 : 12 : 21" }, + }; av_log_set_level(AV_LOG_DEBUG); + for (i = 0; i < FF_ARRAY_ELEMS(fmt_timespec_entries); i++) { + char *p; + struct fmt_timespec_entry *e = &fmt_timespec_entries[i]; + printf("fmt:'%s' spec:'%s' -> ", e->fmt, e->timespec); + p = av_small_strptime(e->timespec, e->fmt, &tm); + if (p) { + printf("%04d-%02d-%2d %02d:%02d:%02d\n", + 1900+tm.tm_year, tm.tm_mon+1, tm.tm_mday, + tm.tm_hour, tm.tm_min, tm.tm_sec); + } else { + printf("error\n"); + } + } +} - for (i = 0; i < FF_ARRAY_ELEMS(color_names); i++) { - if (av_parse_color(rgba, color_names[i], -1, NULL) >= 0) +static void test_av_parse_time(void) +{ + int i; + int64_t tv; + time_t tvi; + struct tm *tm; + static char tzstr[] = "TZ=CET-1"; + static const char * const time_string[] = { + "now", + "12:35:46", + "2000-12-20 0:02:47.5z", + "2012 - 02-22 17:44:07", + "2000-12-20T010247.6", + "2000-12-12 1:35:46+05:30", + "2002-12-12 22:30:40-02", + }; + static const char * const duration_string[] = { + "2:34:56.79", + "-1:23:45.67", + "42.1729", + "-1729.42", + "12:34", + }; + + av_log_set_level(AV_LOG_DEBUG); + putenv(tzstr); + printf("(now is 2012-03-17 09:14:13.2 +0100, local time is UTC+1)\n"); + for (i = 0; i < FF_ARRAY_ELEMS(time_string); i++) { + printf("%-24s -> ", time_string[i]); + if (av_parse_time(&tv, time_string[i], 0)) { + printf("error\n"); + } else { + tvi = tv / 1000000; + tm = gmtime(&tvi); + printf("%14"PRIi64".%06d = %04d-%02d-%02dT%02d:%02d:%02dZ\n", + tv / 1000000, (int)(tv % 1000000), + tm->tm_year + 1900, tm->tm_mon + 1, tm->tm_mday, + tm->tm_hour, tm->tm_min, tm->tm_sec); + } + } + for (i = 0; i < FF_ARRAY_ELEMS(duration_string); i++) { + printf("%-24s -> ", duration_string[i]); + if (av_parse_time(&tv, duration_string[i], 1)) { + printf("error\n"); + } else { + printf("%+21"PRIi64"\n", tv); + } + } +} + +static void test_av_get_known_color_name(void) +{ + int i; + const uint8_t *rgba; + const char *color; + + for (i = 0; i < FF_ARRAY_ELEMS(color_table); ++i) { + color = av_get_known_color_name(i, &rgba); + if (color) printf("%s -> R(%d) G(%d) B(%d) A(%d)\n", - color_names[i], rgba[0], rgba[1], rgba[2], rgba[3]); + color, rgba[0], rgba[1], rgba[2], rgba[3]); + else + printf("Color ID: %d not found\n", i); } +} + +static void test_av_find_info_tag(void) +{ + static const char args[] = "?tag1=val1&tag2=val2&tag3=val3&tag41=value 41&tag42=random1"; + static const char *tags[] = {"tag1", "tag2", "tag3", "tag4", "tag41", "41", "random1"}; + char buff[16]; + int i; + + for (i = 0; i < FF_ARRAY_ELEMS(tags); ++i) { + if (av_find_info_tag(buff, sizeof(buff), tags[i], args)) + printf("%d. %s found: %s\n", i, tags[i], buff); + else + printf("%d. %s not found\n", i, tags[i]); + } +} + +int main(void) +{ + printf("Testing av_parse_video_rate()\n"); + test_av_parse_video_rate(); + + printf("\nTesting av_parse_color()\n"); + test_av_parse_color(); + + printf("\nTesting av_small_strptime()\n"); + test_av_small_strptime(); + + printf("\nTesting av_parse_time()\n"); + test_av_parse_time(); + + printf("\nTesting av_get_known_color_name()\n"); + test_av_get_known_color_name(); + printf("\nTesting av_find_info_tag()\n"); + test_av_find_info_tag(); return 0; } diff --git a/libavutil/parseutils.c b/libavutil/parseutils.c index f4248114b5..a4efd79628 100644 --- a/libavutil/parseutils.c +++ b/libavutil/parseutils.c @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -31,6 +31,37 @@ #include "random_seed.h" #include "time_internal.h" #include "parseutils.h" +#include "time.h" + +#ifdef TEST + +#define av_get_random_seed av_get_random_seed_deterministic +static uint32_t av_get_random_seed_deterministic(void); + +#define av_gettime() 1331972053200000 + +#endif + +int av_parse_ratio(AVRational *q, const char *str, int max, + int log_offset, void *log_ctx) +{ + char c; + int ret; + + if (sscanf(str, "%d:%d%c", &q->num, &q->den, &c) != 2) { + double d; + ret = av_expr_parse_and_eval(&d, str, NULL, NULL, + NULL, NULL, NULL, NULL, + NULL, log_offset, log_ctx); + if (ret < 0) + return ret; + *q = av_d2q(d, max); + } else { + av_reduce(&q->num, &q->den, q->num, q->den, max); + } + + return 0; +} typedef struct VideoSizeAbbr { const char *abbr; @@ -80,8 +111,20 @@ static const VideoSizeAbbr video_size_abbrs[] = { { "hd480", 852, 480 }, { "hd720", 1280, 720 }, { "hd1080", 1920,1080 }, + { "2k", 2048,1080 }, /* Digital Cinema System Specification */ { "2kdci", 2048,1080 }, + { "2kflat", 1998,1080 }, + { "2kscope", 2048, 858 }, + { "4k", 4096,2160 }, /* Digital Cinema System Specification */ { "4kdci", 4096,2160 }, + { "4kflat", 3996,2160 }, + { "4kscope", 4096,1716 }, + { "nhd", 640,360 }, + { "hqvga", 240,160 }, + { "wqvga", 400,240 }, + { "fwqvga", 432,240 }, + { "hvga", 480,320 }, + { "qhd", 960,540 }, { "uhd2160", 3840,2160 }, { "uhd4320", 7680,4320 }, }; @@ -101,7 +144,7 @@ int av_parse_video_size(int *width_ptr, int *height_ptr, const char *str) { int i; int n = FF_ARRAY_ELEMS(video_size_abbrs); - char *p; + const char *p; int width = 0, height = 0; for (i = 0; i < n; i++) { @@ -112,10 +155,14 @@ int av_parse_video_size(int *width_ptr, int *height_ptr, const char *str) } } if (i == n) { - width = strtol(str, &p, 10); + width = strtol(str, (void*)&p, 10); if (*p) p++; - height = strtol(p, &p, 10); + height = strtol(p, (void*)&p, 10); + + /* trailing extraneous data detected, like in 123x345foobar */ + if (*p) + return AVERROR(EINVAL); } if (width <= 0 || height <= 0) return AVERROR(EINVAL); @@ -128,7 +175,6 @@ int av_parse_video_rate(AVRational *rate, const char *arg) { int i, ret; int n = FF_ARRAY_ELEMS(video_rate_abbrs); - double res; /* First, we check our abbreviation table */ for (i = 0; i < n; ++i) @@ -138,10 +184,8 @@ int av_parse_video_rate(AVRational *rate, const char *arg) } /* Then, we try to parse it as fraction */ - if ((ret = av_expr_parse_and_eval(&res, arg, NULL, NULL, NULL, NULL, NULL, NULL, - NULL, 0, NULL)) < 0) + if ((ret = av_parse_ratio_quiet(rate, arg, 1001000)) < 0) return ret; - *rate = av_d2q(res, 1001000); if (rate->num <= 0 || rate->den <= 0) return AVERROR(EINVAL); return 0; @@ -152,7 +196,7 @@ typedef struct ColorEntry { uint8_t rgb_color[3]; ///< RGB values for the color } ColorEntry; -static ColorEntry color_table[] = { +static const ColorEntry color_table[] = { { "AliceBlue", { 0xF0, 0xF8, 0xFF } }, { "AntiqueWhite", { 0xFA, 0xEB, 0xD7 } }, { "Aqua", { 0x00, 0xFF, 0xFF } }, @@ -220,8 +264,8 @@ static ColorEntry color_table[] = { { "LightCoral", { 0xF0, 0x80, 0x80 } }, { "LightCyan", { 0xE0, 0xFF, 0xFF } }, { "LightGoldenRodYellow", { 0xFA, 0xFA, 0xD2 } }, - { "LightGrey", { 0xD3, 0xD3, 0xD3 } }, { "LightGreen", { 0x90, 0xEE, 0x90 } }, + { "LightGrey", { 0xD3, 0xD3, 0xD3 } }, { "LightPink", { 0xFF, 0xB6, 0xC1 } }, { "LightSalmon", { 0xFF, 0xA0, 0x7A } }, { "LightSeaGreen", { 0x20, 0xB2, 0xAA } }, @@ -364,7 +408,11 @@ int av_parse_color(uint8_t *rgba_color, const char *color_string, int slen, if (!strncmp(alpha_string, "0x", 2)) { alpha = strtoul(alpha_string, &tail, 16); } else { - alpha = 255 * strtod(alpha_string, &tail); + double norm_alpha = strtod(alpha_string, &tail); + if (norm_alpha < 0.0 || norm_alpha > 1.0) + alpha = 256; + else + alpha = 255 * norm_alpha; } if (tail == alpha_string || *tail || alpha > 255 || alpha < 0) { @@ -378,6 +426,20 @@ int av_parse_color(uint8_t *rgba_color, const char *color_string, int slen, return 0; } +const char *av_get_known_color_name(int color_idx, const uint8_t **rgbp) +{ + const ColorEntry *color; + + if ((unsigned)color_idx >= FF_ARRAY_ELEMS(color_table)) + return NULL; + + color = &color_table[color_idx]; + if (rgbp) + *rgbp = color->rgb_color; + + return color->name; +} + /* get a positive number between n_min and n_max, for a maximum length of len_max. Return -1 if error. */ static int date_get_num(const char **pp, @@ -404,7 +466,7 @@ static int date_get_num(const char **pp, return val; } -const char *av_small_strptime(const char *p, const char *fmt, struct tm *dt) +char *av_small_strptime(const char *p, const char *fmt, struct tm *dt) { int c, val; @@ -421,7 +483,9 @@ const char *av_small_strptime(const char *p, const char *fmt, struct tm *dt) c = *fmt++; switch(c) { case 'H': - val = date_get_num(&p, 0, 23, 2); + case 'J': + val = date_get_num(&p, 0, c == 'H' ? 23 : INT_MAX, 2); + if (val == -1) return NULL; dt->tm_hour = val; @@ -470,7 +534,7 @@ const char *av_small_strptime(const char *p, const char *fmt, struct tm *dt) } } - return p; + return (char*)p; } time_t av_timegm(struct tm *tm) @@ -484,7 +548,7 @@ time_t av_timegm(struct tm *tm) y--; } - t = 86400 * + t = 86400LL * (d + (153 * m - 457) / 5 + 365 * y + y / 4 - y / 100 + y / 400 - 719469); t += 3600 * tm->tm_hour + 60 * tm->tm_min + tm->tm_sec; @@ -494,70 +558,64 @@ time_t av_timegm(struct tm *tm) int av_parse_time(int64_t *timeval, const char *timestr, int duration) { - const char *p; - int64_t t; + const char *p, *q; + int64_t t, now64; + time_t now; struct tm dt = { 0 }, tmbuf; + int today = 0, negative = 0, microseconds = 0; int i; static const char * const date_fmt[] = { - "%Y-%m-%d", + "%Y - %m - %d", "%Y%m%d", }; static const char * const time_fmt[] = { "%H:%M:%S", "%H%M%S", }; - const char *q; - int is_utc, len; - char lastch; - int negative = 0; - - time_t now = time(0); - - len = strlen(timestr); - if (len > 0) - lastch = timestr[len - 1]; - else - lastch = '\0'; - is_utc = (lastch == 'z' || lastch == 'Z'); + static const char * const tz_fmt[] = { + "%H:%M", + "%H%M", + "%H", + }; p = timestr; q = NULL; + *timeval = INT64_MIN; if (!duration) { - if (!av_strncasecmp(timestr, "now", len)) { - *timeval = (int64_t) now * 1000000; + now64 = av_gettime(); + now = now64 / 1000000; + + if (!av_strcasecmp(timestr, "now")) { + *timeval = now64; return 0; } /* parse the year-month-day part */ for (i = 0; i < FF_ARRAY_ELEMS(date_fmt); i++) { q = av_small_strptime(p, date_fmt[i], &dt); - if (q) { + if (q) break; - } } /* if the year-month-day part is missing, then take the * current year-month-day time */ if (!q) { - if (is_utc) { - dt = *gmtime_r(&now, &tmbuf); - } else { - dt = *localtime_r(&now, &tmbuf); - } - dt.tm_hour = dt.tm_min = dt.tm_sec = 0; - } else { - p = q; + today = 1; + q = p; } + p = q; - if (*p == 'T' || *p == 't' || *p == ' ') + if (*p == 'T' || *p == 't') p++; + else + while (av_isspace(*p)) + p++; /* parse the hour-minute-second part */ for (i = 0; i < FF_ARRAY_ELEMS(time_fmt); i++) { q = av_small_strptime(p, time_fmt[i], &dt); - if (q) { + if (q) break; - } } } else { /* parse timestr as a duration */ @@ -566,16 +624,18 @@ int av_parse_time(int64_t *timeval, const char *timestr, int duration) ++p; } /* parse timestr as HH:MM:SS */ - q = av_small_strptime(p, time_fmt[0], &dt); + q = av_small_strptime(p, "%J:%M:%S", &dt); + if (!q) { + /* parse timestr as MM:SS */ + q = av_small_strptime(p, "%M:%S", &dt); + dt.tm_hour = 0; + } if (!q) { char *o; /* parse timestr as S+ */ dt.tm_sec = strtol(p, &o, 10); - if (o == p) { - /* the parsing didn't succeed */ - *timeval = INT64_MIN; + if (o == p) /* the parsing didn't succeed */ return AVERROR(EINVAL); - } dt.tm_min = 0; dt.tm_hour = 0; q = o; @@ -583,35 +643,60 @@ int av_parse_time(int64_t *timeval, const char *timestr, int duration) } /* Now we have all the fields that we can get */ - if (!q) { - *timeval = INT64_MIN; + if (!q) return AVERROR(EINVAL); + + /* parse the .m... part */ + if (*q == '.') { + int n; + q++; + for (n = 100000; n >= 1; n /= 10, q++) { + if (!av_isdigit(*q)) + break; + microseconds += n * (*q - '0'); + } + while (av_isdigit(*q)) + q++; } if (duration) { t = dt.tm_hour * 3600 + dt.tm_min * 60 + dt.tm_sec; } else { - dt.tm_isdst = -1; /* unknown */ - if (is_utc) { - t = av_timegm(&dt); - } else { - t = mktime(&dt); + int is_utc = *q == 'Z' || *q == 'z'; + int tzoffset = 0; + q += is_utc; + if (!today && !is_utc && (*q == '+' || *q == '-')) { + struct tm tz = { 0 }; + int sign = (*q == '+' ? -1 : 1); + q++; + p = q; + for (i = 0; i < FF_ARRAY_ELEMS(tz_fmt); i++) { + q = av_small_strptime(p, tz_fmt[i], &tz); + if (q) + break; + } + if (!q) + return AVERROR(EINVAL); + tzoffset = sign * (tz.tm_hour * 60 + tz.tm_min) * 60; + is_utc = 1; } + if (today) { /* fill in today's date */ + struct tm dt2 = is_utc ? *gmtime_r(&now, &tmbuf) : *localtime_r(&now, &tmbuf); + dt2.tm_hour = dt.tm_hour; + dt2.tm_min = dt.tm_min; + dt2.tm_sec = dt.tm_sec; + dt = dt2; + } + t = is_utc ? av_timegm(&dt) : mktime(&dt); + t += tzoffset; } - t *= 1000000; + /* Check that we are at the end of the string */ + if (*q) + return AVERROR(EINVAL); - /* parse the .m... part */ - if (*q == '.') { - int val, n; - q++; - for (val = 0, n = 100000; n >= 1; n /= 10, q++) { - if (!av_isdigit(*q)) - break; - val += n * (*q - '0'); - } - t += val; - } + t *= 1000000; + t += microseconds; *timeval = negative ? -t : t; return 0; } diff --git a/libavutil/parseutils.h b/libavutil/parseutils.h index 8e99634773..e66d24b76e 100644 --- a/libavutil/parseutils.h +++ b/libavutil/parseutils.h @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -29,6 +29,30 @@ */ /** + * Parse str and store the parsed ratio in q. + * + * Note that a ratio with infinite (1/0) or negative value is + * considered valid, so you should check on the returned value if you + * want to exclude those values. + * + * The undefined value can be expressed using the "0:0" string. + * + * @param[in,out] q pointer to the AVRational which will contain the ratio + * @param[in] str the string to parse: it has to be a string in the format + * num:den, a float number or an expression + * @param[in] max the maximum allowed numerator and denominator + * @param[in] log_offset log level offset which is applied to the log + * level of log_ctx + * @param[in] log_ctx parent logging context + * @return >= 0 on success, a negative error code otherwise + */ +int av_parse_ratio(AVRational *q, const char *str, int max, + int log_offset, void *log_ctx); + +#define av_parse_ratio_quiet(rate, str, max) \ + av_parse_ratio(rate, str, max, AV_LOG_MAX_OFFSET, NULL) + +/** * Parse str and put in width_ptr and height_ptr the detected values. * * @param[in,out] width_ptr pointer to the variable which will contain the detected @@ -75,6 +99,19 @@ int av_parse_color(uint8_t *rgba_color, const char *color_string, int slen, void *log_ctx); /** + * Get the name of a color from the internal table of hard-coded named + * colors. + * + * This function is meant to enumerate the color names recognized by + * av_parse_color(). + * + * @param color_idx index of the requested color, starting from 0 + * @param rgbp if not NULL, will point to a 3-elements array with the color value in RGB + * @return the color name string or NULL if color_idx is not in the array + */ +const char *av_get_known_color_name(int color_idx, const uint8_t **rgb); + +/** * Parse timestr and return in *time a corresponding number of * microseconds. * @@ -88,7 +125,7 @@ int av_parse_color(uint8_t *rgba_color, const char *color_string, int slen, * @param timestr a string representing a date or a duration. * - If a date the syntax is: * @code - * [{YYYY-MM-DD|YYYYMMDD}[T|t| ]]{{HH[:MM[:SS[.m...]]]}|{HH[MM[SS[.m...]]]}}[Z] + * [{YYYY-MM-DD|YYYYMMDD}[T|t| ]]{{HH:MM:SS[.m...]]]}|{HHMMSS[.m...]]]}}[Z] * now * @endcode * If the value is "now" it takes the current time. @@ -98,12 +135,12 @@ int av_parse_color(uint8_t *rgba_color, const char *color_string, int slen, * year-month-day. * - If a duration the syntax is: * @code - * [-]HH[:MM[:SS[.m...]]] + * [-][HH:]MM:SS[.m...] * [-]S+[.m...] * @endcode * @param duration flag which tells how to interpret timestr, if not * zero timestr is interpreted as a duration, otherwise as a date - * @return 0 in case of success, a negative value corresponding to an + * @return >= 0 in case of success, a negative value corresponding to an * AVERROR code otherwise */ int av_parse_time(int64_t *timeval, const char *timestr, int duration); @@ -121,12 +158,13 @@ int av_find_info_tag(char *arg, int arg_size, const char *tag1, const char *info * * Parse the input string p according to the format string fmt and * store its results in the structure dt. - * - * Neither text and locale's alternative representation are supported. + * This implementation supports only a subset of the formats supported + * by the standard strptime(). * * The supported input field descriptors are listed below. * - %H: the hour as a decimal number, using a 24-hour clock, in the * range '00' through '23' + * - %J: hours as a decimal number, in the range '0' through INT_MAX * - %M: the minute as a decimal number, using a 24-hour clock, in the * range '00' through '59' * - %S: the second as a decimal number, using a 24-hour clock, in the @@ -145,7 +183,7 @@ int av_find_info_tag(char *arg, int arg_size, const char *tag1, const char *info * is consumed the return value points to the null byte at the end of * the string. On failure NULL is returned. */ -const char *av_small_strptime(const char *p, const char *fmt, struct tm *dt); +char *av_small_strptime(const char *p, const char *fmt, struct tm *dt); /** * Convert the decomposed UTC time in tm to a time_t value. diff --git a/libavutil/pca-test.c b/libavutil/pca-test.c new file mode 100644 index 0000000000..e56d6185b1 --- /dev/null +++ b/libavutil/pca-test.c @@ -0,0 +1,102 @@ +/* + * principal component analysis (PCA) + * Copyright (c) 2004 Michael Niedermayer <michaelni@gmx.at> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "pca.c" +#include "lfg.h" + +#undef printf +#include <stdio.h> +#include <stdlib.h> + +int main(void){ + PCA *pca; + int i, j, k; +#define LEN 8 + double eigenvector[LEN*LEN]; + double eigenvalue[LEN]; + AVLFG prng; + + av_lfg_init(&prng, 1); + + pca= ff_pca_init(LEN); + + for(i=0; i<9000000; i++){ + double v[2*LEN+100]; + double sum=0; + int pos = av_lfg_get(&prng) % LEN; + int v2 = av_lfg_get(&prng) % 101 - 50; + v[0] = av_lfg_get(&prng) % 101 - 50; + for(j=1; j<8; j++){ + if(j<=pos) v[j]= v[0]; + else v[j]= v2; + sum += v[j]; + } +/* for(j=0; j<LEN; j++){ + v[j] -= v[pos]; + }*/ +// sum += av_lfg_get(&prng) % 10; +/* for(j=0; j<LEN; j++){ + v[j] -= sum/LEN; + }*/ +// lbt1(v+100,v+100,LEN); + ff_pca_add(pca, v); + } + + + ff_pca(pca, eigenvector, eigenvalue); + for(i=0; i<LEN; i++){ + pca->count= 1; + pca->mean[i]= 0; + +// (0.5^|x|)^2 = 0.5^2|x| = 0.25^|x| + + +// pca.covariance[i + i*LEN]= pow(0.5, fabs + for(j=i; j<LEN; j++){ + printf("%f ", pca->covariance[i + j*LEN]); + } + printf("\n"); + } + + for(i=0; i<LEN; i++){ + double v[LEN]; + double error=0; + memset(v, 0, sizeof(v)); + for(j=0; j<LEN; j++){ + for(k=0; k<LEN; k++){ + v[j] += pca->covariance[FFMIN(k,j) + FFMAX(k,j)*LEN] * eigenvector[i + k*LEN]; + } + v[j] /= eigenvalue[i]; + error += fabs(v[j] - eigenvector[i + j*LEN]); + } + printf("%f ", error); + } + printf("\n"); + + for(i=0; i<LEN; i++){ + for(j=0; j<LEN; j++){ + printf("%9.6f ", eigenvector[i + j*LEN]); + } + printf(" %9.1f %f\n", eigenvalue[i], eigenvalue[i]/eigenvalue[0]); + } + + return 0; +} diff --git a/libavutil/pca.c b/libavutil/pca.c new file mode 100644 index 0000000000..4e52c7b362 --- /dev/null +++ b/libavutil/pca.c @@ -0,0 +1,173 @@ +/* + * principal component analysis (PCA) + * Copyright (c) 2004 Michael Niedermayer <michaelni@gmx.at> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +/** + * @file + * principal component analysis (PCA) + */ + +#include "common.h" +#include "pca.h" + +typedef struct PCA{ + int count; + int n; + double *covariance; + double *mean; + double *z; +}PCA; + +PCA *ff_pca_init(int n){ + PCA *pca; + if(n<=0) + return NULL; + + pca= av_mallocz(sizeof(*pca)); + if (!pca) + return NULL; + + pca->n= n; + pca->z = av_malloc_array(n, sizeof(*pca->z)); + pca->count=0; + pca->covariance= av_calloc(n*n, sizeof(double)); + pca->mean= av_calloc(n, sizeof(double)); + + if (!pca->z || !pca->covariance || !pca->mean) { + ff_pca_free(pca); + return NULL; + } + + return pca; +} + +void ff_pca_free(PCA *pca){ + av_freep(&pca->covariance); + av_freep(&pca->mean); + av_freep(&pca->z); + av_free(pca); +} + +void ff_pca_add(PCA *pca, const double *v){ + int i, j; + const int n= pca->n; + + for(i=0; i<n; i++){ + pca->mean[i] += v[i]; + for(j=i; j<n; j++) + pca->covariance[j + i*n] += v[i]*v[j]; + } + pca->count++; +} + +int ff_pca(PCA *pca, double *eigenvector, double *eigenvalue){ + int i, j, pass; + int k=0; + const int n= pca->n; + double *z = pca->z; + + memset(eigenvector, 0, sizeof(double)*n*n); + + for(j=0; j<n; j++){ + pca->mean[j] /= pca->count; + eigenvector[j + j*n] = 1.0; + for(i=0; i<=j; i++){ + pca->covariance[j + i*n] /= pca->count; + pca->covariance[j + i*n] -= pca->mean[i] * pca->mean[j]; + pca->covariance[i + j*n] = pca->covariance[j + i*n]; + } + eigenvalue[j]= pca->covariance[j + j*n]; + z[j]= 0; + } + + for(pass=0; pass < 50; pass++){ + double sum=0; + + for(i=0; i<n; i++) + for(j=i+1; j<n; j++) + sum += fabs(pca->covariance[j + i*n]); + + if(sum == 0){ + for(i=0; i<n; i++){ + double maxvalue= -1; + for(j=i; j<n; j++){ + if(eigenvalue[j] > maxvalue){ + maxvalue= eigenvalue[j]; + k= j; + } + } + eigenvalue[k]= eigenvalue[i]; + eigenvalue[i]= maxvalue; + for(j=0; j<n; j++){ + double tmp= eigenvector[k + j*n]; + eigenvector[k + j*n]= eigenvector[i + j*n]; + eigenvector[i + j*n]= tmp; + } + } + return pass; + } + + for(i=0; i<n; i++){ + for(j=i+1; j<n; j++){ + double covar= pca->covariance[j + i*n]; + double t,c,s,tau,theta, h; + + if(pass < 3 && fabs(covar) < sum / (5*n*n)) //FIXME why pass < 3 + continue; + if(fabs(covar) == 0.0) //FIXME should not be needed + continue; + if(pass >=3 && fabs((eigenvalue[j]+z[j])/covar) > (1LL<<32) && fabs((eigenvalue[i]+z[i])/covar) > (1LL<<32)){ + pca->covariance[j + i*n]=0.0; + continue; + } + + h= (eigenvalue[j]+z[j]) - (eigenvalue[i]+z[i]); + theta=0.5*h/covar; + t=1.0/(fabs(theta)+sqrt(1.0+theta*theta)); + if(theta < 0.0) t = -t; + + c=1.0/sqrt(1+t*t); + s=t*c; + tau=s/(1.0+c); + z[i] -= t*covar; + z[j] += t*covar; + +#define ROTATE(a,i,j,k,l) {\ + double g=a[j + i*n];\ + double h=a[l + k*n];\ + a[j + i*n]=g-s*(h+g*tau);\ + a[l + k*n]=h+s*(g-h*tau); } + for(k=0; k<n; k++) { + if(k!=i && k!=j){ + ROTATE(pca->covariance,FFMIN(k,i),FFMAX(k,i),FFMIN(k,j),FFMAX(k,j)) + } + ROTATE(eigenvector,k,i,k,j) + } + pca->covariance[j + i*n]=0.0; + } + } + for (i=0; i<n; i++) { + eigenvalue[i] += z[i]; + z[i]=0.0; + } + } + + return -1; +} diff --git a/libavutil/pca.h b/libavutil/pca.h new file mode 100644 index 0000000000..992bb2eb25 --- /dev/null +++ b/libavutil/pca.h @@ -0,0 +1,35 @@ +/* + * principal component analysis (PCA) + * Copyright (c) 2004 Michael Niedermayer <michaelni@gmx.at> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +/** + * @file + * principal component analysis (PCA) + */ + +#ifndef AVUTIL_PCA_H +#define AVUTIL_PCA_H + +struct PCA *ff_pca_init(int n); +void ff_pca_free(struct PCA *pca); +void ff_pca_add(struct PCA *pca, const double *v); +int ff_pca(struct PCA *pca, double *eigenvector, double *eigenvalue); + +#endif /* AVUTIL_PCA_H */ diff --git a/libavutil/pixdesc-test.c b/libavutil/pixdesc-test.c new file mode 100644 index 0000000000..cd7d4c5f7b --- /dev/null +++ b/libavutil/pixdesc-test.c @@ -0,0 +1,46 @@ +/* + * pixel format descriptor + * Copyright (c) 2009 Michael Niedermayer <michaelni@gmx.at> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "pixdesc.c" + +int main(void){ + int i; + int err=0; + int skip = 0; + + for (i=0; i<AV_PIX_FMT_NB*2; i++) { + const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(i); + if(!desc || !desc->name) { + skip ++; + continue; + } + if (skip) { + av_log(NULL, AV_LOG_INFO, "%3d unused pixel format values\n", skip); + skip = 0; + } + av_log(NULL, AV_LOG_INFO, "pix fmt %s avg_bpp:%d colortype:%d\n", desc->name, av_get_padded_bits_per_pixel(desc), get_color_type(desc)); + if ((!(desc->flags & AV_PIX_FMT_FLAG_ALPHA)) != (desc->nb_components != 2 && desc->nb_components != 4)) { + av_log(NULL, AV_LOG_ERROR, "Alpha flag mismatch\n"); + err = 1; + } + } + return err; +} diff --git a/libavutil/pixdesc.c b/libavutil/pixdesc.c index b383cd4fb5..c630d840ce 100644 --- a/libavutil/pixdesc.c +++ b/libavutil/pixdesc.c @@ -2,26 +2,27 @@ * pixel format descriptor * Copyright (c) 2009 Michael Niedermayer <michaelni@gmx.at> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ #include <stdio.h> #include <string.h> +#include "avassert.h" #include "avstring.h" #include "common.h" #include "pixfmt.h" @@ -236,6 +237,18 @@ static const AVPixFmtDescriptor av_pix_fmt_descriptors[AV_PIX_FMT_NB] = { }, .flags = AV_PIX_FMT_FLAG_PLANAR, }, + [AV_PIX_FMT_YUVJ411P] = { + .name = "yuvj411p", + .nb_components = 3, + .log2_chroma_w = 2, + .log2_chroma_h = 0, + .comp = { + { 0, 1, 0, 0, 8, 0, 7, 1 }, /* Y */ + { 1, 1, 0, 0, 8, 0, 7, 1 }, /* U */ + { 2, 1, 0, 0, 8, 0, 7, 1 }, /* V */ + }, + .flags = AV_PIX_FMT_FLAG_PLANAR, + }, [AV_PIX_FMT_GRAY8] = { .name = "gray", .nb_components = 1, @@ -323,6 +336,12 @@ static const AVPixFmtDescriptor av_pix_fmt_descriptors[AV_PIX_FMT_NB] = { .flags = AV_PIX_FMT_FLAG_HWACCEL, }, #endif /* FF_API_XVMC */ +#if !FF_API_XVMC + [AV_PIX_FMT_XVMC] = { + .name = "xvmc", + .flags = AV_PIX_FMT_FLAG_HWACCEL, + }, +#endif /* !FF_API_XVMC */ [AV_PIX_FMT_UYVY422] = { .name = "uyvy422", .nb_components = 3, @@ -493,6 +512,54 @@ static const AVPixFmtDescriptor av_pix_fmt_descriptors[AV_PIX_FMT_NB] = { }, .flags = AV_PIX_FMT_FLAG_RGB | AV_PIX_FMT_FLAG_ALPHA, }, + [AV_PIX_FMT_0RGB] = { + .name = "0rgb", + .nb_components= 3, + .log2_chroma_w= 0, + .log2_chroma_h= 0, + .comp = { + { 0, 4, 1, 0, 8, 3, 7, 2 }, /* R */ + { 0, 4, 2, 0, 8, 3, 7, 3 }, /* G */ + { 0, 4, 3, 0, 8, 3, 7, 4 }, /* B */ + }, + .flags = AV_PIX_FMT_FLAG_RGB, + }, + [AV_PIX_FMT_RGB0] = { + .name = "rgb0", + .nb_components= 3, + .log2_chroma_w= 0, + .log2_chroma_h= 0, + .comp = { + { 0, 4, 0, 0, 8, 3, 7, 1 }, /* R */ + { 0, 4, 1, 0, 8, 3, 7, 2 }, /* G */ + { 0, 4, 2, 0, 8, 3, 7, 3 }, /* B */ + }, + .flags = AV_PIX_FMT_FLAG_RGB, + }, + [AV_PIX_FMT_0BGR] = { + .name = "0bgr", + .nb_components= 3, + .log2_chroma_w= 0, + .log2_chroma_h= 0, + .comp = { + { 0, 4, 3, 0, 8, 3, 7, 4 }, /* R */ + { 0, 4, 2, 0, 8, 3, 7, 3 }, /* G */ + { 0, 4, 1, 0, 8, 3, 7, 2 }, /* B */ + }, + .flags = AV_PIX_FMT_FLAG_RGB, + }, + [AV_PIX_FMT_BGR0] = { + .name = "bgr0", + .nb_components= 3, + .log2_chroma_w= 0, + .log2_chroma_h= 0, + .comp = { + { 0, 4, 2, 0, 8, 3, 7, 3 }, /* R */ + { 0, 4, 1, 0, 8, 3, 7, 2 }, /* G */ + { 0, 4, 0, 0, 8, 3, 7, 1 }, /* B */ + }, + .flags = AV_PIX_FMT_FLAG_RGB, + }, [AV_PIX_FMT_GRAY16BE] = { .name = "gray16be", .nb_components = 1, @@ -538,6 +605,54 @@ static const AVPixFmtDescriptor av_pix_fmt_descriptors[AV_PIX_FMT_NB] = { }, .flags = AV_PIX_FMT_FLAG_PLANAR, }, + [AV_PIX_FMT_YUV440P10LE] = { + .name = "yuv440p10le", + .nb_components = 3, + .log2_chroma_w = 0, + .log2_chroma_h = 1, + .comp = { + { 0, 2, 0, 0, 10, 1, 9, 1 }, /* Y */ + { 1, 2, 0, 0, 10, 1, 9, 1 }, /* U */ + { 2, 2, 0, 0, 10, 1, 9, 1 }, /* V */ + }, + .flags = AV_PIX_FMT_FLAG_PLANAR, + }, + [AV_PIX_FMT_YUV440P10BE] = { + .name = "yuv440p10be", + .nb_components = 3, + .log2_chroma_w = 0, + .log2_chroma_h = 1, + .comp = { + { 0, 2, 0, 0, 10, 1, 9, 1 }, /* Y */ + { 1, 2, 0, 0, 10, 1, 9, 1 }, /* U */ + { 2, 2, 0, 0, 10, 1, 9, 1 }, /* V */ + }, + .flags = AV_PIX_FMT_FLAG_BE | AV_PIX_FMT_FLAG_PLANAR, + }, + [AV_PIX_FMT_YUV440P12LE] = { + .name = "yuv440p12le", + .nb_components = 3, + .log2_chroma_w = 0, + .log2_chroma_h = 1, + .comp = { + { 0, 2, 0, 0, 12, 1, 11, 1 }, /* Y */ + { 1, 2, 0, 0, 12, 1, 11, 1 }, /* U */ + { 2, 2, 0, 0, 12, 1, 11, 1 }, /* V */ + }, + .flags = AV_PIX_FMT_FLAG_PLANAR, + }, + [AV_PIX_FMT_YUV440P12BE] = { + .name = "yuv440p12be", + .nb_components = 3, + .log2_chroma_w = 0, + .log2_chroma_h = 1, + .comp = { + { 0, 2, 0, 0, 12, 1, 11, 1 }, /* Y */ + { 1, 2, 0, 0, 12, 1, 11, 1 }, /* U */ + { 2, 2, 0, 0, 12, 1, 11, 1 }, /* V */ + }, + .flags = AV_PIX_FMT_FLAG_BE | AV_PIX_FMT_FLAG_PLANAR, + }, [AV_PIX_FMT_YUVA420P] = { .name = "yuva420p", .nb_components = 4, @@ -551,7 +666,7 @@ static const AVPixFmtDescriptor av_pix_fmt_descriptors[AV_PIX_FMT_NB] = { }, .flags = AV_PIX_FMT_FLAG_PLANAR | AV_PIX_FMT_FLAG_ALPHA, }, - [AV_PIX_FMT_YUVA422P] = { + [AV_PIX_FMT_YUVA422P] = { .name = "yuva422p", .nb_components = 4, .log2_chroma_w = 1, @@ -1120,12 +1235,6 @@ static const AVPixFmtDescriptor av_pix_fmt_descriptors[AV_PIX_FMT_NB] = { .flags = AV_PIX_FMT_FLAG_HWACCEL, }, #endif - [AV_PIX_FMT_VDA_VLD] = { - .name = "vda_vld", - .log2_chroma_w = 1, - .log2_chroma_h = 1, - .flags = AV_PIX_FMT_FLAG_HWACCEL, - }, [AV_PIX_FMT_YUV420P9LE] = { .name = "yuv420p9le", .nb_components = 3, @@ -1174,6 +1283,54 @@ static const AVPixFmtDescriptor av_pix_fmt_descriptors[AV_PIX_FMT_NB] = { }, .flags = AV_PIX_FMT_FLAG_BE | AV_PIX_FMT_FLAG_PLANAR, }, + [AV_PIX_FMT_YUV420P12LE] = { + .name = "yuv420p12le", + .nb_components = 3, + .log2_chroma_w = 1, + .log2_chroma_h = 1, + .comp = { + { 0, 2, 0, 0, 12, 1, 11, 1 }, /* Y */ + { 1, 2, 0, 0, 12, 1, 11, 1 }, /* U */ + { 2, 2, 0, 0, 12, 1, 11, 1 }, /* V */ + }, + .flags = AV_PIX_FMT_FLAG_PLANAR, + }, + [AV_PIX_FMT_YUV420P12BE] = { + .name = "yuv420p12be", + .nb_components = 3, + .log2_chroma_w = 1, + .log2_chroma_h = 1, + .comp = { + { 0, 2, 0, 0, 12, 1, 11, 1 }, /* Y */ + { 1, 2, 0, 0, 12, 1, 11, 1 }, /* U */ + { 2, 2, 0, 0, 12, 1, 11, 1 }, /* V */ + }, + .flags = AV_PIX_FMT_FLAG_BE | AV_PIX_FMT_FLAG_PLANAR, + }, + [AV_PIX_FMT_YUV420P14LE] = { + .name = "yuv420p14le", + .nb_components = 3, + .log2_chroma_w = 1, + .log2_chroma_h = 1, + .comp = { + { 0, 2, 0, 0, 14, 1, 13, 1 }, /* Y */ + { 1, 2, 0, 0, 14, 1, 13, 1 }, /* U */ + { 2, 2, 0, 0, 14, 1, 13, 1 }, /* V */ + }, + .flags = AV_PIX_FMT_FLAG_PLANAR, + }, + [AV_PIX_FMT_YUV420P14BE] = { + .name = "yuv420p14be", + .nb_components = 3, + .log2_chroma_w = 1, + .log2_chroma_h = 1, + .comp = { + { 0, 2, 0, 0, 14, 1, 13, 1 }, /* Y */ + { 1, 2, 0, 0, 14, 1, 13, 1 }, /* U */ + { 2, 2, 0, 0, 14, 1, 13, 1 }, /* V */ + }, + .flags = AV_PIX_FMT_FLAG_BE | AV_PIX_FMT_FLAG_PLANAR, + }, [AV_PIX_FMT_YUV420P16LE] = { .name = "yuv420p16le", .nb_components = 3, @@ -1246,6 +1403,54 @@ static const AVPixFmtDescriptor av_pix_fmt_descriptors[AV_PIX_FMT_NB] = { }, .flags = AV_PIX_FMT_FLAG_BE | AV_PIX_FMT_FLAG_PLANAR, }, + [AV_PIX_FMT_YUV422P12LE] = { + .name = "yuv422p12le", + .nb_components = 3, + .log2_chroma_w = 1, + .log2_chroma_h = 0, + .comp = { + { 0, 2, 0, 0, 12, 1, 11, 1 }, /* Y */ + { 1, 2, 0, 0, 12, 1, 11, 1 }, /* U */ + { 2, 2, 0, 0, 12, 1, 11, 1 }, /* V */ + }, + .flags = AV_PIX_FMT_FLAG_PLANAR, + }, + [AV_PIX_FMT_YUV422P12BE] = { + .name = "yuv422p12be", + .nb_components = 3, + .log2_chroma_w = 1, + .log2_chroma_h = 0, + .comp = { + { 0, 2, 0, 0, 12, 1, 11, 1 }, /* Y */ + { 1, 2, 0, 0, 12, 1, 11, 1 }, /* U */ + { 2, 2, 0, 0, 12, 1, 11, 1 }, /* V */ + }, + .flags = AV_PIX_FMT_FLAG_BE | AV_PIX_FMT_FLAG_PLANAR, + }, + [AV_PIX_FMT_YUV422P14LE] = { + .name = "yuv422p14le", + .nb_components = 3, + .log2_chroma_w = 1, + .log2_chroma_h = 0, + .comp = { + { 0, 2, 0, 0, 14, 1, 13, 1 }, /* Y */ + { 1, 2, 0, 0, 14, 1, 13, 1 }, /* U */ + { 2, 2, 0, 0, 14, 1, 13, 1 }, /* V */ + }, + .flags = AV_PIX_FMT_FLAG_PLANAR, + }, + [AV_PIX_FMT_YUV422P14BE] = { + .name = "yuv422p14be", + .nb_components = 3, + .log2_chroma_w = 1, + .log2_chroma_h = 0, + .comp = { + { 0, 2, 0, 0, 14, 1, 13, 1 }, /* Y */ + { 1, 2, 0, 0, 14, 1, 13, 1 }, /* U */ + { 2, 2, 0, 0, 14, 1, 13, 1 }, /* V */ + }, + .flags = AV_PIX_FMT_FLAG_BE | AV_PIX_FMT_FLAG_PLANAR, + }, [AV_PIX_FMT_YUV422P16LE] = { .name = "yuv422p16le", .nb_components = 3, @@ -1342,6 +1547,54 @@ static const AVPixFmtDescriptor av_pix_fmt_descriptors[AV_PIX_FMT_NB] = { }, .flags = AV_PIX_FMT_FLAG_BE | AV_PIX_FMT_FLAG_PLANAR, }, + [AV_PIX_FMT_YUV444P12LE] = { + .name = "yuv444p12le", + .nb_components = 3, + .log2_chroma_w = 0, + .log2_chroma_h = 0, + .comp = { + { 0, 2, 0, 0, 12, 1, 11, 1 }, /* Y */ + { 1, 2, 0, 0, 12, 1, 11, 1 }, /* U */ + { 2, 2, 0, 0, 12, 1, 11, 1 }, /* V */ + }, + .flags = AV_PIX_FMT_FLAG_PLANAR, + }, + [AV_PIX_FMT_YUV444P12BE] = { + .name = "yuv444p12be", + .nb_components = 3, + .log2_chroma_w = 0, + .log2_chroma_h = 0, + .comp = { + { 0, 2, 0, 0, 12, 1, 11, 1 }, /* Y */ + { 1, 2, 0, 0, 12, 1, 11, 1 }, /* U */ + { 2, 2, 0, 0, 12, 1, 11, 1 }, /* V */ + }, + .flags = AV_PIX_FMT_FLAG_BE | AV_PIX_FMT_FLAG_PLANAR, + }, + [AV_PIX_FMT_YUV444P14LE] = { + .name = "yuv444p14le", + .nb_components = 3, + .log2_chroma_w = 0, + .log2_chroma_h = 0, + .comp = { + { 0, 2, 0, 0, 14, 1, 13, 1 }, /* Y */ + { 1, 2, 0, 0, 14, 1, 13, 1 }, /* U */ + { 2, 2, 0, 0, 14, 1, 13, 1 }, /* V */ + }, + .flags = AV_PIX_FMT_FLAG_PLANAR, + }, + [AV_PIX_FMT_YUV444P14BE] = { + .name = "yuv444p14be", + .nb_components = 3, + .log2_chroma_w = 0, + .log2_chroma_h = 0, + .comp = { + { 0, 2, 0, 0, 14, 1, 13, 1 }, /* Y */ + { 1, 2, 0, 0, 14, 1, 13, 1 }, /* U */ + { 2, 2, 0, 0, 14, 1, 13, 1 }, /* V */ + }, + .flags = AV_PIX_FMT_FLAG_BE | AV_PIX_FMT_FLAG_PLANAR, + }, [AV_PIX_FMT_D3D11VA_VLD] = { .name = "d3d11va_vld", .log2_chroma_w = 1, @@ -1354,6 +1607,12 @@ static const AVPixFmtDescriptor av_pix_fmt_descriptors[AV_PIX_FMT_NB] = { .log2_chroma_h = 1, .flags = AV_PIX_FMT_FLAG_HWACCEL, }, + [AV_PIX_FMT_VDA_VLD] = { + .name = "vda_vld", + .log2_chroma_w = 1, + .log2_chroma_h = 1, + .flags = AV_PIX_FMT_FLAG_HWACCEL, + }, [AV_PIX_FMT_YA8] = { .name = "ya8", .nb_components = 2, @@ -1382,6 +1641,10 @@ static const AVPixFmtDescriptor av_pix_fmt_descriptors[AV_PIX_FMT_NB] = { }, .flags = AV_PIX_FMT_FLAG_BE | AV_PIX_FMT_FLAG_ALPHA, }, + [AV_PIX_FMT_VIDEOTOOLBOX] = { + .name = "videotoolbox_vld", + .flags = AV_PIX_FMT_FLAG_HWACCEL, + }, [AV_PIX_FMT_GBRP] = { .name = "gbrp", .nb_components = 3, @@ -1442,6 +1705,54 @@ static const AVPixFmtDescriptor av_pix_fmt_descriptors[AV_PIX_FMT_NB] = { }, .flags = AV_PIX_FMT_FLAG_BE | AV_PIX_FMT_FLAG_PLANAR | AV_PIX_FMT_FLAG_RGB, }, + [AV_PIX_FMT_GBRP12LE] = { + .name = "gbrp12le", + .nb_components = 3, + .log2_chroma_w = 0, + .log2_chroma_h = 0, + .comp = { + { 2, 2, 0, 0, 12, 1, 11, 1 }, /* R */ + { 0, 2, 0, 0, 12, 1, 11, 1 }, /* G */ + { 1, 2, 0, 0, 12, 1, 11, 1 }, /* B */ + }, + .flags = AV_PIX_FMT_FLAG_PLANAR | AV_PIX_FMT_FLAG_RGB, + }, + [AV_PIX_FMT_GBRP12BE] = { + .name = "gbrp12be", + .nb_components = 3, + .log2_chroma_w = 0, + .log2_chroma_h = 0, + .comp = { + { 2, 2, 0, 0, 12, 1, 11, 1 }, /* R */ + { 0, 2, 0, 0, 12, 1, 11, 1 }, /* G */ + { 1, 2, 0, 0, 12, 1, 11, 1 }, /* B */ + }, + .flags = AV_PIX_FMT_FLAG_BE | AV_PIX_FMT_FLAG_PLANAR | AV_PIX_FMT_FLAG_RGB, + }, + [AV_PIX_FMT_GBRP14LE] = { + .name = "gbrp14le", + .nb_components = 3, + .log2_chroma_w = 0, + .log2_chroma_h = 0, + .comp = { + { 2, 2, 0, 0, 14, 1, 13, 1 }, /* R */ + { 0, 2, 0, 0, 14, 1, 13, 1 }, /* G */ + { 1, 2, 0, 0, 14, 1, 13, 1 }, /* B */ + }, + .flags = AV_PIX_FMT_FLAG_PLANAR | AV_PIX_FMT_FLAG_RGB, + }, + [AV_PIX_FMT_GBRP14BE] = { + .name = "gbrp14be", + .nb_components = 3, + .log2_chroma_w = 0, + .log2_chroma_h = 0, + .comp = { + { 2, 2, 0, 0, 14, 1, 13, 1 }, /* R */ + { 0, 2, 0, 0, 14, 1, 13, 1 }, /* G */ + { 1, 2, 0, 0, 14, 1, 13, 1 }, /* B */ + }, + .flags = AV_PIX_FMT_FLAG_BE | AV_PIX_FMT_FLAG_PLANAR | AV_PIX_FMT_FLAG_RGB, + }, [AV_PIX_FMT_GBRP16LE] = { .name = "gbrp16le", .nb_components = 3, @@ -1538,6 +1849,87 @@ static const AVPixFmtDescriptor av_pix_fmt_descriptors[AV_PIX_FMT_NB] = { }, .flags = AV_PIX_FMT_FLAG_BE, }, + +#define BAYER8_DESC_COMMON \ + .nb_components= 3, \ + .log2_chroma_w= 0, \ + .log2_chroma_h= 0, \ + .comp = { \ + {0,1,0,0,2,0,1,1},\ + {0,1,0,0,4,0,3,1},\ + {0,1,0,0,2,0,1,1},\ + }, \ + +#define BAYER16_DESC_COMMON \ + .nb_components= 3, \ + .log2_chroma_w= 0, \ + .log2_chroma_h= 0, \ + .comp = { \ + {0,2,0,0,4,1,3,1},\ + {0,2,0,0,8,1,7,1},\ + {0,2,0,0,4,1,3,1},\ + }, \ + + [AV_PIX_FMT_BAYER_BGGR8] = { + .name = "bayer_bggr8", + BAYER8_DESC_COMMON + .flags = AV_PIX_FMT_FLAG_RGB, + }, + [AV_PIX_FMT_BAYER_BGGR16LE] = { + .name = "bayer_bggr16le", + BAYER16_DESC_COMMON + .flags = AV_PIX_FMT_FLAG_RGB, + }, + [AV_PIX_FMT_BAYER_BGGR16BE] = { + .name = "bayer_bggr16be", + BAYER16_DESC_COMMON + .flags = AV_PIX_FMT_FLAG_BE | AV_PIX_FMT_FLAG_RGB, + }, + [AV_PIX_FMT_BAYER_RGGB8] = { + .name = "bayer_rggb8", + BAYER8_DESC_COMMON + .flags = AV_PIX_FMT_FLAG_RGB, + }, + [AV_PIX_FMT_BAYER_RGGB16LE] = { + .name = "bayer_rggb16le", + BAYER16_DESC_COMMON + .flags = AV_PIX_FMT_FLAG_RGB, + }, + [AV_PIX_FMT_BAYER_RGGB16BE] = { + .name = "bayer_rggb16be", + BAYER16_DESC_COMMON + .flags = AV_PIX_FMT_FLAG_BE | AV_PIX_FMT_FLAG_RGB, + }, + [AV_PIX_FMT_BAYER_GBRG8] = { + .name = "bayer_gbrg8", + BAYER8_DESC_COMMON + .flags = AV_PIX_FMT_FLAG_RGB, + }, + [AV_PIX_FMT_BAYER_GBRG16LE] = { + .name = "bayer_gbrg16le", + BAYER16_DESC_COMMON + .flags = AV_PIX_FMT_FLAG_RGB, + }, + [AV_PIX_FMT_BAYER_GBRG16BE] = { + .name = "bayer_gbrg16be", + BAYER16_DESC_COMMON + .flags = AV_PIX_FMT_FLAG_BE | AV_PIX_FMT_FLAG_RGB, + }, + [AV_PIX_FMT_BAYER_GRBG8] = { + .name = "bayer_grbg8", + BAYER8_DESC_COMMON + .flags = AV_PIX_FMT_FLAG_RGB, + }, + [AV_PIX_FMT_BAYER_GRBG16LE] = { + .name = "bayer_grbg16le", + BAYER16_DESC_COMMON + .flags = AV_PIX_FMT_FLAG_RGB, + }, + [AV_PIX_FMT_BAYER_GRBG16BE] = { + .name = "bayer_grbg16be", + BAYER16_DESC_COMMON + .flags = AV_PIX_FMT_FLAG_BE | AV_PIX_FMT_FLAG_RGB, + }, [AV_PIX_FMT_NV16] = { .name = "nv16", .nb_components = 3, @@ -1590,6 +1982,84 @@ static const AVPixFmtDescriptor av_pix_fmt_descriptors[AV_PIX_FMT_NB] = { .name = "cuda", .flags = AV_PIX_FMT_FLAG_HWACCEL, }, + [AV_PIX_FMT_AYUV64LE] = { + .name = "ayuv64le", + .nb_components = 4, + .log2_chroma_w = 0, + .log2_chroma_h = 0, + .comp = { + { 0, 8, 2, 0, 16, 7, 15, 3 }, /* Y */ + { 0, 8, 4, 0, 16, 7, 15, 5 }, /* U */ + { 0, 8, 6, 0, 16, 7, 15, 7 }, /* V */ + { 0, 8, 0, 0, 16, 7, 15, 1 }, /* A */ + }, + .flags = AV_PIX_FMT_FLAG_ALPHA, + }, + [AV_PIX_FMT_AYUV64BE] = { + .name = "ayuv64be", + .nb_components = 4, + .log2_chroma_w = 0, + .log2_chroma_h = 0, + .comp = { + { 0, 8, 2, 0, 16, 7, 15, 3 }, /* Y */ + { 0, 8, 4, 0, 16, 7, 15, 5 }, /* U */ + { 0, 8, 6, 0, 16, 7, 15, 7 }, /* V */ + { 0, 8, 0, 0, 16, 7, 15, 1 }, /* A */ + }, + .flags = AV_PIX_FMT_FLAG_BE | AV_PIX_FMT_FLAG_ALPHA, + }, + [AV_PIX_FMT_P010LE] = { + .name = "p010le", + .nb_components = 3, + .log2_chroma_w = 1, + .log2_chroma_h = 1, + .comp = { + { 0, 2, 0, 6, 10, 1, 9, 1 }, /* Y */ + { 1, 4, 0, 6, 10, 3, 9, 1 }, /* U */ + { 1, 4, 2, 6, 10, 3, 9, 3 }, /* V */ + }, + .flags = AV_PIX_FMT_FLAG_PLANAR, + }, + [AV_PIX_FMT_P010BE] = { + .name = "p010be", + .nb_components = 3, + .log2_chroma_w = 1, + .log2_chroma_h = 1, + .comp = { + { 0, 2, 0, 6, 10, 1, 9, 1 }, /* Y */ + { 1, 4, 0, 6, 10, 3, 9, 1 }, /* U */ + { 1, 4, 2, 6, 10, 3, 9, 3 }, /* V */ + }, + .flags = AV_PIX_FMT_FLAG_PLANAR | AV_PIX_FMT_FLAG_BE, + }, + [AV_PIX_FMT_GBRAP12LE] = { + .name = "gbrap12le", + .nb_components = 4, + .log2_chroma_w = 0, + .log2_chroma_h = 0, + .comp = { + { 2, 2, 0, 0, 12, 1, 11, 1 }, /* R */ + { 0, 2, 0, 0, 12, 1, 11, 1 }, /* G */ + { 1, 2, 0, 0, 12, 1, 11, 1 }, /* B */ + { 3, 2, 0, 0, 12, 1, 11, 1 }, /* A */ + }, + .flags = AV_PIX_FMT_FLAG_PLANAR | AV_PIX_FMT_FLAG_RGB | + AV_PIX_FMT_FLAG_ALPHA, + }, + [AV_PIX_FMT_GBRAP12BE] = { + .name = "gbrap12be", + .nb_components = 4, + .log2_chroma_w = 0, + .log2_chroma_h = 0, + .comp = { + { 2, 2, 0, 0, 12, 1, 11, 1 }, /* R */ + { 0, 2, 0, 0, 12, 1, 11, 1 }, /* G */ + { 1, 2, 0, 0, 12, 1, 11, 1 }, /* B */ + { 3, 2, 0, 0, 12, 1, 11, 1 }, /* A */ + }, + .flags = AV_PIX_FMT_FLAG_BE | AV_PIX_FMT_FLAG_PLANAR | + AV_PIX_FMT_FLAG_RGB | AV_PIX_FMT_FLAG_ALPHA, + }, }; #if FF_API_PLUS1_MINUS1 FF_ENABLE_DEPRECATION_WARNINGS @@ -1602,14 +2072,14 @@ static const char *color_range_names[AVCOL_RANGE_NB] = { static const char *color_primaries_names[AVCOL_PRI_NB] = { "reserved", "bt709", "unknown", "reserved", "bt470m", "bt470bg", "smpte170m", "smpte240m", "film", "bt2020", - "smptest428-1", + "smpte428-1", }; static const char *color_transfer_names[AVCOL_TRC_NB] = { "reserved", "bt709", "unknown", "reserved", "bt470m", "bt470bg", "smpte170m", "smpte240m", "linear", "log100", "log316", "iec61966-2-4", "bt1361e", "iec61966-2-1", - "bt2020-10", "bt2020-20", "smptest2084", "smptest428-1", + "bt2020-10", "bt2020-20", "smpte2084", "smpte428-1", }; static const char *color_space_names[AVCOL_SPC_NB] = { @@ -1685,6 +2155,26 @@ int av_get_bits_per_pixel(const AVPixFmtDescriptor *pixdesc) return bits >> log2_pixels; } +int av_get_padded_bits_per_pixel(const AVPixFmtDescriptor *pixdesc) +{ + int c, bits = 0; + int log2_pixels = pixdesc->log2_chroma_w + pixdesc->log2_chroma_h; + int steps[4] = {0}; + + for (c = 0; c < pixdesc->nb_components; c++) { + const AVComponentDescriptor *comp = &pixdesc->comp[c]; + int s = c == 1 || c == 2 ? 0 : log2_pixels; + steps[comp->plane] = comp->step << s; + } + for (c = 0; c < 4; c++) + bits += steps[c]; + + if(!(pixdesc->flags & AV_PIX_FMT_FLAG_BITSTREAM)) + bits *= 8; + + return bits >> log2_pixels; +} + char *av_get_pix_fmt_string(char *buf, int buf_size, enum AVPixelFormat pix_fmt) { @@ -1711,8 +2201,11 @@ const AVPixFmtDescriptor *av_pix_fmt_desc_next(const AVPixFmtDescriptor *prev) { if (!prev) return &av_pix_fmt_descriptors[0]; - if (prev - av_pix_fmt_descriptors < FF_ARRAY_ELEMS(av_pix_fmt_descriptors) - 1) - return prev + 1; + while (prev - av_pix_fmt_descriptors < FF_ARRAY_ELEMS(av_pix_fmt_descriptors) - 1) { + prev++; + if (prev->name) + return prev; + } return NULL; } @@ -1752,56 +2245,262 @@ int av_pix_fmt_count_planes(enum AVPixelFormat pix_fmt) return ret; } +void ff_check_pixfmt_descriptors(void){ + int i, j; + + for (i=0; i<FF_ARRAY_ELEMS(av_pix_fmt_descriptors); i++) { + const AVPixFmtDescriptor *d = &av_pix_fmt_descriptors[i]; + uint8_t fill[4][8+6+3] = {{0}}; + uint8_t *data[4] = {fill[0], fill[1], fill[2], fill[3]}; + int linesize[4] = {0,0,0,0}; + uint16_t tmp[2]; + + if (!d->name && !d->nb_components && !d->log2_chroma_w && !d->log2_chroma_h && !d->flags) + continue; +// av_log(NULL, AV_LOG_DEBUG, "Checking: %s\n", d->name); + av_assert0(d->log2_chroma_w <= 3); + av_assert0(d->log2_chroma_h <= 3); + av_assert0(d->nb_components <= 4); + av_assert0(d->name && d->name[0]); + av_assert0((d->nb_components==4 || d->nb_components==2) == !!(d->flags & AV_PIX_FMT_FLAG_ALPHA)); + av_assert2(av_get_pix_fmt(d->name) == i); + + for (j=0; j<FF_ARRAY_ELEMS(d->comp); j++) { + const AVComponentDescriptor *c = &d->comp[j]; + if(j>=d->nb_components) { + av_assert0(!c->plane && !c->step && !c->offset && !c->shift && !c->depth); + continue; + } + if (d->flags & AV_PIX_FMT_FLAG_BITSTREAM) { + av_assert0(c->step >= c->depth); + } else { + av_assert0(8*c->step >= c->depth); + } + if (!strncmp(d->name, "bayer_", 6)) + continue; + av_read_image_line(tmp, (void*)data, linesize, d, 0, 0, j, 2, 0); + av_assert0(tmp[0] == 0 && tmp[1] == 0); + tmp[0] = tmp[1] = (1<<c->depth) - 1; + av_write_image_line(tmp, data, linesize, d, 0, 0, j, 2); + } + } +} + enum AVPixelFormat av_pix_fmt_swap_endianness(enum AVPixelFormat pix_fmt) { -#define PIX_FMT_SWAP_ENDIANNESS(fmt) \ - case AV_PIX_FMT_ ## fmt ## BE: return AV_PIX_FMT_ ## fmt ## LE; \ - case AV_PIX_FMT_ ## fmt ## LE: return AV_PIX_FMT_ ## fmt ## BE - - switch (pix_fmt) { - PIX_FMT_SWAP_ENDIANNESS(GRAY16); - PIX_FMT_SWAP_ENDIANNESS(YA16); - PIX_FMT_SWAP_ENDIANNESS(RGB48); - PIX_FMT_SWAP_ENDIANNESS(RGB565); - PIX_FMT_SWAP_ENDIANNESS(RGB555); - PIX_FMT_SWAP_ENDIANNESS(RGB444); - PIX_FMT_SWAP_ENDIANNESS(BGR48); - PIX_FMT_SWAP_ENDIANNESS(BGR565); - PIX_FMT_SWAP_ENDIANNESS(BGR555); - PIX_FMT_SWAP_ENDIANNESS(BGR444); - - PIX_FMT_SWAP_ENDIANNESS(YUV420P9); - PIX_FMT_SWAP_ENDIANNESS(YUV422P9); - PIX_FMT_SWAP_ENDIANNESS(YUV444P9); - PIX_FMT_SWAP_ENDIANNESS(YUV420P10); - PIX_FMT_SWAP_ENDIANNESS(YUV422P10); - PIX_FMT_SWAP_ENDIANNESS(YUV444P10); - PIX_FMT_SWAP_ENDIANNESS(YUV420P16); - PIX_FMT_SWAP_ENDIANNESS(YUV422P16); - PIX_FMT_SWAP_ENDIANNESS(YUV444P16); - - PIX_FMT_SWAP_ENDIANNESS(GBRP9); - PIX_FMT_SWAP_ENDIANNESS(GBRP10); - PIX_FMT_SWAP_ENDIANNESS(GBRP16); - PIX_FMT_SWAP_ENDIANNESS(YUVA420P9); - PIX_FMT_SWAP_ENDIANNESS(YUVA422P9); - PIX_FMT_SWAP_ENDIANNESS(YUVA444P9); - PIX_FMT_SWAP_ENDIANNESS(YUVA420P10); - PIX_FMT_SWAP_ENDIANNESS(YUVA422P10); - PIX_FMT_SWAP_ENDIANNESS(YUVA444P10); - PIX_FMT_SWAP_ENDIANNESS(YUVA420P16); - PIX_FMT_SWAP_ENDIANNESS(YUVA422P16); - PIX_FMT_SWAP_ENDIANNESS(YUVA444P16); - - PIX_FMT_SWAP_ENDIANNESS(XYZ12); - PIX_FMT_SWAP_ENDIANNESS(NV20); - PIX_FMT_SWAP_ENDIANNESS(RGBA64); - PIX_FMT_SWAP_ENDIANNESS(BGRA64); - default: + const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt); + char name[16]; + int i; + + if (!desc || strlen(desc->name) < 2) + return AV_PIX_FMT_NONE; + av_strlcpy(name, desc->name, sizeof(name)); + i = strlen(name) - 2; + if (strcmp(name + i, "be") && strcmp(name + i, "le")) return AV_PIX_FMT_NONE; + + name[i] ^= 'b' ^ 'l'; + + return get_pix_fmt_internal(name); +} + +#define FF_COLOR_NA -1 +#define FF_COLOR_RGB 0 /**< RGB color space */ +#define FF_COLOR_GRAY 1 /**< gray color space */ +#define FF_COLOR_YUV 2 /**< YUV color space. 16 <= Y <= 235, 16 <= U, V <= 240 */ +#define FF_COLOR_YUV_JPEG 3 /**< YUV color space. 0 <= Y <= 255, 0 <= U, V <= 255 */ +#define FF_COLOR_XYZ 4 + +#define pixdesc_has_alpha(pixdesc) \ + ((pixdesc)->nb_components == 2 || (pixdesc)->nb_components == 4 || (pixdesc)->flags & AV_PIX_FMT_FLAG_PAL) + + +static int get_color_type(const AVPixFmtDescriptor *desc) { + if (desc->flags & AV_PIX_FMT_FLAG_PAL) + return FF_COLOR_RGB; + + if(desc->nb_components == 1 || desc->nb_components == 2) + return FF_COLOR_GRAY; + + if(desc->name && !strncmp(desc->name, "yuvj", 4)) + return FF_COLOR_YUV_JPEG; + + if(desc->name && !strncmp(desc->name, "xyz", 3)) + return FF_COLOR_XYZ; + + if(desc->flags & AV_PIX_FMT_FLAG_RGB) + return FF_COLOR_RGB; + + if(desc->nb_components == 0) + return FF_COLOR_NA; + + return FF_COLOR_YUV; +} + +static int get_pix_fmt_depth(int *min, int *max, enum AVPixelFormat pix_fmt) +{ + const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(pix_fmt); + int i; + + if (!desc || !desc->nb_components) { + *min = *max = 0; + return AVERROR(EINVAL); + } + + *min = INT_MAX, *max = -INT_MAX; + for (i = 0; i < desc->nb_components; i++) { + *min = FFMIN(desc->comp[i].depth, *min); + *max = FFMAX(desc->comp[i].depth, *max); + } + return 0; +} + +static int get_pix_fmt_score(enum AVPixelFormat dst_pix_fmt, + enum AVPixelFormat src_pix_fmt, + unsigned *lossp, unsigned consider) +{ + const AVPixFmtDescriptor *src_desc = av_pix_fmt_desc_get(src_pix_fmt); + const AVPixFmtDescriptor *dst_desc = av_pix_fmt_desc_get(dst_pix_fmt); + int src_color, dst_color; + int src_min_depth, src_max_depth, dst_min_depth, dst_max_depth; + int ret, loss, i, nb_components; + int score = INT_MAX - 1; + + if (dst_pix_fmt >= AV_PIX_FMT_NB || dst_pix_fmt <= AV_PIX_FMT_NONE) + return ~0; + + /* compute loss */ + *lossp = loss = 0; + + if (dst_pix_fmt == src_pix_fmt) + return INT_MAX; + + if ((ret = get_pix_fmt_depth(&src_min_depth, &src_max_depth, src_pix_fmt)) < 0) + return ret; + if ((ret = get_pix_fmt_depth(&dst_min_depth, &dst_max_depth, dst_pix_fmt)) < 0) + return ret; + + src_color = get_color_type(src_desc); + dst_color = get_color_type(dst_desc); + if (dst_pix_fmt == AV_PIX_FMT_PAL8) + nb_components = FFMIN(src_desc->nb_components, 4); + else + nb_components = FFMIN(src_desc->nb_components, dst_desc->nb_components); + + for (i = 0; i < nb_components; i++) { + int depth_minus1 = (dst_pix_fmt == AV_PIX_FMT_PAL8) ? 7/nb_components : (dst_desc->comp[i].depth - 1); + if (src_desc->comp[i].depth - 1 > depth_minus1 && (consider & FF_LOSS_DEPTH)) { + loss |= FF_LOSS_DEPTH; + score -= 65536 >> depth_minus1; + } + } + + if (consider & FF_LOSS_RESOLUTION) { + if (dst_desc->log2_chroma_w > src_desc->log2_chroma_w) { + loss |= FF_LOSS_RESOLUTION; + score -= 256 << dst_desc->log2_chroma_w; + } + if (dst_desc->log2_chroma_h > src_desc->log2_chroma_h) { + loss |= FF_LOSS_RESOLUTION; + score -= 256 << dst_desc->log2_chroma_h; + } + // don't favor 422 over 420 if downsampling is needed, because 420 has much better support on the decoder side + if (dst_desc->log2_chroma_w == 1 && src_desc->log2_chroma_w == 0 && + dst_desc->log2_chroma_h == 1 && src_desc->log2_chroma_h == 0 ) { + score += 512; + } + } + + if(consider & FF_LOSS_COLORSPACE) + switch(dst_color) { + case FF_COLOR_RGB: + if (src_color != FF_COLOR_RGB && + src_color != FF_COLOR_GRAY) + loss |= FF_LOSS_COLORSPACE; + break; + case FF_COLOR_GRAY: + if (src_color != FF_COLOR_GRAY) + loss |= FF_LOSS_COLORSPACE; + break; + case FF_COLOR_YUV: + if (src_color != FF_COLOR_YUV) + loss |= FF_LOSS_COLORSPACE; + break; + case FF_COLOR_YUV_JPEG: + if (src_color != FF_COLOR_YUV_JPEG && + src_color != FF_COLOR_YUV && + src_color != FF_COLOR_GRAY) + loss |= FF_LOSS_COLORSPACE; + break; + default: + /* fail safe test */ + if (src_color != dst_color) + loss |= FF_LOSS_COLORSPACE; + break; + } + if(loss & FF_LOSS_COLORSPACE) + score -= (nb_components * 65536) >> FFMIN(dst_desc->comp[0].depth - 1, src_desc->comp[0].depth - 1); + + if (dst_color == FF_COLOR_GRAY && + src_color != FF_COLOR_GRAY && (consider & FF_LOSS_CHROMA)) { + loss |= FF_LOSS_CHROMA; + score -= 2 * 65536; + } + if (!pixdesc_has_alpha(dst_desc) && (pixdesc_has_alpha(src_desc) && (consider & FF_LOSS_ALPHA))) { + loss |= FF_LOSS_ALPHA; + score -= 65536; + } + if (dst_pix_fmt == AV_PIX_FMT_PAL8 && (consider & FF_LOSS_COLORQUANT) && + (src_pix_fmt != AV_PIX_FMT_PAL8 && (src_color != FF_COLOR_GRAY || (pixdesc_has_alpha(src_desc) && (consider & FF_LOSS_ALPHA))))) { + loss |= FF_LOSS_COLORQUANT; + score -= 65536; + } + + *lossp = loss; + return score; +} + +int av_get_pix_fmt_loss(enum AVPixelFormat dst_pix_fmt, + enum AVPixelFormat src_pix_fmt, + int has_alpha) +{ + int loss; + int ret = get_pix_fmt_score(dst_pix_fmt, src_pix_fmt, &loss, has_alpha ? ~0 : ~FF_LOSS_ALPHA); + if (ret < 0) + return ret; + return loss; +} + +enum AVPixelFormat av_find_best_pix_fmt_of_2(enum AVPixelFormat dst_pix_fmt1, enum AVPixelFormat dst_pix_fmt2, + enum AVPixelFormat src_pix_fmt, int has_alpha, int *loss_ptr) +{ + enum AVPixelFormat dst_pix_fmt; + int loss1, loss2, loss_mask; + const AVPixFmtDescriptor *desc1 = av_pix_fmt_desc_get(dst_pix_fmt1); + const AVPixFmtDescriptor *desc2 = av_pix_fmt_desc_get(dst_pix_fmt2); + int score1, score2; + + loss_mask= loss_ptr?~*loss_ptr:~0; /* use loss mask if provided */ + if(!has_alpha) + loss_mask &= ~FF_LOSS_ALPHA; + + score1 = get_pix_fmt_score(dst_pix_fmt1, src_pix_fmt, &loss1, loss_mask); + score2 = get_pix_fmt_score(dst_pix_fmt2, src_pix_fmt, &loss2, loss_mask); + + if (score1 == score2) { + if(av_get_padded_bits_per_pixel(desc2) != av_get_padded_bits_per_pixel(desc1)) { + dst_pix_fmt = av_get_padded_bits_per_pixel(desc2) < av_get_padded_bits_per_pixel(desc1) ? dst_pix_fmt2 : dst_pix_fmt1; + } else { + dst_pix_fmt = desc2->nb_components < desc1->nb_components ? dst_pix_fmt2 : dst_pix_fmt1; + } + } else { + dst_pix_fmt = score1 < score2 ? dst_pix_fmt2 : dst_pix_fmt1; } -#undef PIX_FMT_SWAP_ENDIANNESS + + if (loss_ptr) + *loss_ptr = av_get_pix_fmt_loss(dst_pix_fmt, src_pix_fmt, has_alpha); + return dst_pix_fmt; } const char *av_color_range_name(enum AVColorRange range) @@ -1833,4 +2532,3 @@ const char *av_chroma_location_name(enum AVChromaLocation location) return (unsigned) location < AVCHROMA_LOC_NB ? chroma_location_names[location] : NULL; } - diff --git a/libavutil/pixdesc.h b/libavutil/pixdesc.h index 3bb10f777a..3b0bcdb3d8 100644 --- a/libavutil/pixdesc.h +++ b/libavutil/pixdesc.h @@ -2,20 +2,20 @@ * pixel format descriptor * Copyright (c) 2009 Michael Niedermayer <michaelni@gmx.at> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -26,6 +26,7 @@ #include "attributes.h" #include "pixfmt.h" +#include "version.h" typedef struct AVComponentDescriptor { /** @@ -105,14 +106,11 @@ typedef struct AVPixFmtDescriptor { uint64_t flags; /** - * Parameters that describe how pixels are packed. If the format - * has chroma components, they must be stored in comp[1] and - * comp[2]. - * If the format is RGB-like, the first component is R, followed - * by G and B. - * - * If the format is YUV-like, the first component is Y, followed - * by U and V. + * Parameters that describe how pixels are packed. + * If the format has 1 or 2 components, then luma is 0. + * If the format has 3 or 4 components: + * if the RGB flag is set then 0 is red, 1 is green and 2 is blue; + * otherwise 0 is luma, 1 is chroma-U and 2 is chroma-V. * * If present, the Alpha channel is always the last component. */ @@ -148,14 +146,29 @@ typedef struct AVPixFmtDescriptor { * The pixel format contains RGB-like data (as opposed to YUV/grayscale). */ #define AV_PIX_FMT_FLAG_RGB (1 << 5) + /** - * The pixel format is "pseudo-paletted". This means that Libav treats it as - * paletted internally, but the palette is generated by the decoder and is not - * stored in the file. + * The pixel format is "pseudo-paletted". This means that it contains a + * fixed palette in the 2nd plane but the palette is fixed/constant for each + * PIX_FMT. This allows interpreting the data as if it was PAL8, which can + * in some cases be simpler. Or the data can be interpreted purely based on + * the pixel format without using the palette. + * An example of a pseudo-paletted format is AV_PIX_FMT_GRAY8 */ #define AV_PIX_FMT_FLAG_PSEUDOPAL (1 << 6) + /** - * The pixel format has an alpha channel. + * The pixel format has an alpha channel. This is set on all formats that + * support alpha in some way. The exception is AV_PIX_FMT_PAL8, which can + * carry alpha as part of the palette. Details are explained in the + * AVPixelFormat enum, and are also encoded in the corresponding + * AVPixFmtDescriptor. + * + * The alpha is always straight, never pre-multiplied. + * + * If a codec or a filter does not support alpha, it should set all alpha to + * opaque, or use the equivalent pixel formats without alpha component, e.g. + * AV_PIX_FMT_RGB0 (or AV_PIX_FMT_RGB24 etc.) instead of AV_PIX_FMT_RGBA. */ #define AV_PIX_FMT_FLAG_ALPHA (1 << 7) @@ -206,7 +219,7 @@ void av_write_image_line(const uint16_t *src, uint8_t *data[4], * For example in a little-endian system, first looks for "gray16", * then for "gray16le". * - * Finally if no pixel format has been found, returns PIX_FMT_NONE. + * Finally if no pixel format has been found, returns AV_PIX_FMT_NONE. */ enum AVPixelFormat av_get_pix_fmt(const char *name); @@ -220,7 +233,7 @@ const char *av_get_pix_fmt_name(enum AVPixelFormat pix_fmt); /** * Print in buf the string corresponding to the pixel format with - * number pix_fmt, or an header if pix_fmt is negative. + * number pix_fmt, or a header if pix_fmt is negative. * * @param buf the buffer where to write the string * @param buf_size the size of buf @@ -243,6 +256,12 @@ char *av_get_pix_fmt_string(char *buf, int buf_size, int av_get_bits_per_pixel(const AVPixFmtDescriptor *pixdesc); /** + * Return the number of bits per pixel for the pixel format + * described by pixdesc, including any padding or unused bits. + */ +int av_get_padded_bits_per_pixel(const AVPixFmtDescriptor *pixdesc); + +/** * @return a pixel format descriptor for provided pixel format or NULL if * this pixel format is unknown. */ @@ -267,6 +286,11 @@ enum AVPixelFormat av_pix_fmt_desc_get_id(const AVPixFmtDescriptor *desc); * Utility function to access log2_chroma_w log2_chroma_h from * the pixel format AVPixFmtDescriptor. * + * See av_get_chroma_sub_sample() for a function that asserts a + * valid pixel format instead of returning an error code. + * Its recommended that you use avcodec_get_chroma_sub_sample unless + * you do check the return code! + * * @param[in] pix_fmt the pixel format * @param[out] h_shift store log2_chroma_w (horizontal/width shift) * @param[out] v_shift store log2_chroma_h (vertical/height shift) @@ -292,6 +316,56 @@ int av_pix_fmt_count_planes(enum AVPixelFormat pix_fmt); */ enum AVPixelFormat av_pix_fmt_swap_endianness(enum AVPixelFormat pix_fmt); +#define FF_LOSS_RESOLUTION 0x0001 /**< loss due to resolution change */ +#define FF_LOSS_DEPTH 0x0002 /**< loss due to color depth change */ +#define FF_LOSS_COLORSPACE 0x0004 /**< loss due to color space conversion */ +#define FF_LOSS_ALPHA 0x0008 /**< loss of alpha bits */ +#define FF_LOSS_COLORQUANT 0x0010 /**< loss due to color quantization */ +#define FF_LOSS_CHROMA 0x0020 /**< loss of chroma (e.g. RGB to gray conversion) */ + +/** + * Compute what kind of losses will occur when converting from one specific + * pixel format to another. + * When converting from one pixel format to another, information loss may occur. + * For example, when converting from RGB24 to GRAY, the color information will + * be lost. Similarly, other losses occur when converting from some formats to + * other formats. These losses can involve loss of chroma, but also loss of + * resolution, loss of color depth, loss due to the color space conversion, loss + * of the alpha bits or loss due to color quantization. + * av_get_fix_fmt_loss() informs you about the various types of losses + * which will occur when converting from one pixel format to another. + * + * @param[in] dst_pix_fmt destination pixel format + * @param[in] src_pix_fmt source pixel format + * @param[in] has_alpha Whether the source pixel format alpha channel is used. + * @return Combination of flags informing you what kind of losses will occur + * (maximum loss for an invalid dst_pix_fmt). + */ +int av_get_pix_fmt_loss(enum AVPixelFormat dst_pix_fmt, + enum AVPixelFormat src_pix_fmt, + int has_alpha); + +/** + * Compute what kind of losses will occur when converting from one specific + * pixel format to another. + * When converting from one pixel format to another, information loss may occur. + * For example, when converting from RGB24 to GRAY, the color information will + * be lost. Similarly, other losses occur when converting from some formats to + * other formats. These losses can involve loss of chroma, but also loss of + * resolution, loss of color depth, loss due to the color space conversion, loss + * of the alpha bits or loss due to color quantization. + * av_get_fix_fmt_loss() informs you about the various types of losses + * which will occur when converting from one pixel format to another. + * + * @param[in] dst_pix_fmt destination pixel format + * @param[in] src_pix_fmt source pixel format + * @param[in] has_alpha Whether the source pixel format alpha channel is used. + * @return Combination of flags informing you what kind of losses will occur + * (maximum loss for an invalid dst_pix_fmt). + */ +enum AVPixelFormat av_find_best_pix_fmt_of_2(enum AVPixelFormat dst_pix_fmt1, enum AVPixelFormat dst_pix_fmt2, + enum AVPixelFormat src_pix_fmt, int has_alpha, int *loss_ptr); + /** * @return the name for provided color range or NULL if unknown. */ diff --git a/libavutil/pixelutils-test.c b/libavutil/pixelutils-test.c new file mode 100644 index 0000000000..eebc524c87 --- /dev/null +++ b/libavutil/pixelutils-test.c @@ -0,0 +1,152 @@ +/* + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "pixelutils.c" + +#define W1 320 +#define H1 240 +#define W2 640 +#define H2 480 + +static int run_single_test(const char *test, + const uint8_t *block1, ptrdiff_t stride1, + const uint8_t *block2, ptrdiff_t stride2, + int align, int n) +{ + int out, ref; + av_pixelutils_sad_fn f_ref = sad_c[n - 1]; + av_pixelutils_sad_fn f_out = av_pixelutils_get_sad_fn(n, n, align, NULL); + + switch (align) { + case 0: block1++; block2++; break; + case 1: block2++; break; + case 2: break; + } + + out = f_out(block1, stride1, block2, stride2); + ref = f_ref(block1, stride1, block2, stride2); + printf("[%s] [%c%c] SAD [%s] %dx%d=%d ref=%d\n", + out == ref ? "OK" : "FAIL", + align ? 'A' : 'U', align == 2 ? 'A' : 'U', + test, 1<<n, 1<<n, out, ref); + return out != ref; +} + +static int run_test(const char *test, + const uint8_t *b1, const uint8_t *b2) +{ + int i, a, ret = 0; + + for (a = 0; a < 3; a++) { + const uint8_t *block1 = b1; + const uint8_t *block2 = b2; + + switch (a) { + case 0: block1++; block2++; break; + case 1: block2++; break; + case 2: break; + } + for (i = 1; i <= FF_ARRAY_ELEMS(sad_c); i++) { + int r = run_single_test(test, b1, W1, b2, W2, a, i); + if (r) + ret = r; + } + } + return ret; +} + +int main(void) +{ + int i, align, ret; + uint8_t *buf1 = av_malloc(W1*H1); + uint8_t *buf2 = av_malloc(W2*H2); + uint32_t state = 0; + + if (!buf1 || !buf2) { + fprintf(stderr, "malloc failure\n"); + ret = 1; + goto end; + } + + ff_check_pixfmt_descriptors(); + +#define RANDOM_INIT(buf, size) do { \ + int k; \ + for (k = 0; k < size; k++) { \ + state = state * 1664525 + 1013904223; \ + buf[k] = state>>24; \ + } \ +} while (0) + + /* Normal test with different strides */ + RANDOM_INIT(buf1, W1*H1); + RANDOM_INIT(buf2, W2*H2); + ret = run_test("random", buf1, buf2); + if (ret < 0) + goto end; + + /* Check for maximum SAD */ + memset(buf1, 0xff, W1*H1); + memset(buf2, 0x00, W2*H2); + ret = run_test("max", buf1, buf2); + if (ret < 0) + goto end; + + /* Check for minimum SAD */ + memset(buf1, 0x90, W1*H1); + memset(buf2, 0x90, W2*H2); + ret = run_test("min", buf1, buf2); + if (ret < 0) + goto end; + + /* Exact buffer sizes, to check for overreads */ + for (i = 1; i <= 4; i++) { + for (align = 0; align < 3; align++) { + int size1, size2; + + av_freep(&buf1); + av_freep(&buf2); + + size1 = size2 = 1 << (i << 1); + + switch (align) { + case 0: size1++; size2++; break; + case 1: size2++; break; + case 2: break; + } + + buf1 = av_malloc(size1); + buf2 = av_malloc(size2); + if (!buf1 || !buf2) { + fprintf(stderr, "malloc failure\n"); + ret = 1; + goto end; + } + RANDOM_INIT(buf1, size1); + RANDOM_INIT(buf2, size2); + ret = run_single_test("small", buf1, 1<<i, buf2, 1<<i, align, i); + if (ret < 0) + goto end; + } + } + +end: + av_free(buf1); + av_free(buf2); + return ret; +} diff --git a/libavutil/pixelutils.c b/libavutil/pixelutils.c new file mode 100644 index 0000000000..b663027b42 --- /dev/null +++ b/libavutil/pixelutils.c @@ -0,0 +1,87 @@ +/* + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "config.h" +#include "common.h" +#include "pixelutils.h" +#include "internal.h" + +#if CONFIG_PIXELUTILS + +#include "x86/pixelutils.h" + +static av_always_inline int sad_wxh(const uint8_t *src1, ptrdiff_t stride1, + const uint8_t *src2, ptrdiff_t stride2, + int w, int h) +{ + int x, y, sum = 0; + + for (y = 0; y < h; y++) { + for (x = 0; x < w; x++) + sum += abs(src1[x] - src2[x]); + src1 += stride1; + src2 += stride2; + } + return sum; +} + +#define DECLARE_BLOCK_FUNCTIONS(size) \ +static int block_sad_##size##x##size##_c(const uint8_t *src1, ptrdiff_t stride1, \ + const uint8_t *src2, ptrdiff_t stride2) \ +{ \ + return sad_wxh(src1, stride1, src2, stride2, size, size); \ +} + +DECLARE_BLOCK_FUNCTIONS(2) +DECLARE_BLOCK_FUNCTIONS(4) +DECLARE_BLOCK_FUNCTIONS(8) +DECLARE_BLOCK_FUNCTIONS(16) + +static const av_pixelutils_sad_fn sad_c[] = { + block_sad_2x2_c, + block_sad_4x4_c, + block_sad_8x8_c, + block_sad_16x16_c, +}; + +#endif /* CONFIG_PIXELUTILS */ + +av_pixelutils_sad_fn av_pixelutils_get_sad_fn(int w_bits, int h_bits, int aligned, void *log_ctx) +{ +#if !CONFIG_PIXELUTILS + av_log(log_ctx, AV_LOG_ERROR, "pixelutils support is required " + "but libavutil is not compiled with it\n"); + return NULL; +#else + av_pixelutils_sad_fn sad[FF_ARRAY_ELEMS(sad_c)]; + + memcpy(sad, sad_c, sizeof(sad)); + + if (w_bits < 1 || w_bits > FF_ARRAY_ELEMS(sad) || + h_bits < 1 || h_bits > FF_ARRAY_ELEMS(sad)) + return NULL; + if (w_bits != h_bits) // only squared sad for now + return NULL; + +#if ARCH_X86 + ff_pixelutils_sad_init_x86(sad, aligned); +#endif + + return sad[w_bits - 1]; +#endif +} diff --git a/libavutil/pixelutils.h b/libavutil/pixelutils.h new file mode 100644 index 0000000000..a8dbc157e1 --- /dev/null +++ b/libavutil/pixelutils.h @@ -0,0 +1,52 @@ +/* + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#ifndef AVUTIL_PIXELUTILS_H +#define AVUTIL_PIXELUTILS_H + +#include <stddef.h> +#include <stdint.h> +#include "common.h" + +/** + * Sum of abs(src1[x] - src2[x]) + */ +typedef int (*av_pixelutils_sad_fn)(const uint8_t *src1, ptrdiff_t stride1, + const uint8_t *src2, ptrdiff_t stride2); + +/** + * Get a potentially optimized pointer to a Sum-of-absolute-differences + * function (see the av_pixelutils_sad_fn prototype). + * + * @param w_bits 1<<w_bits is the requested width of the block size + * @param h_bits 1<<h_bits is the requested height of the block size + * @param aligned If set to 2, the returned sad function will assume src1 and + * src2 addresses are aligned on the block size. + * If set to 1, the returned sad function will assume src1 is + * aligned on the block size. + * If set to 0, the returned sad function assume no particular + * alignment. + * @param log_ctx context used for logging, can be NULL + * + * @return a pointer to the SAD function or NULL in case of error (because of + * invalid parameters) + */ +av_pixelutils_sad_fn av_pixelutils_get_sad_fn(int w_bits, int h_bits, + int aligned, void *log_ctx); + +#endif /* AVUTIL_PIXELUTILS_H */ diff --git a/libavutil/pixfmt.h b/libavutil/pixfmt.h index c903a3f649..546eb44440 100644 --- a/libavutil/pixfmt.h +++ b/libavutil/pixfmt.h @@ -1,20 +1,20 @@ /* * copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -29,6 +29,9 @@ #include "libavutil/avconfig.h" #include "version.h" +#define AVPALETTE_SIZE 1024 +#define AVPALETTE_COUNT 256 + /** * Pixel format. * @@ -40,11 +43,11 @@ * big-endian CPUs. * * @par - * When the pixel format is palettized RGB (AV_PIX_FMT_PAL8), the palettized + * When the pixel format is palettized RGB32 (AV_PIX_FMT_PAL8), the palettized * image data is stored in AVFrame.data[0]. The palette is transported in * AVFrame.data[1], is 1024 bytes long (256 4-byte entries) and is * formatted the same as in AV_PIX_FMT_RGB32 described above (i.e., it is - * also endian-specific). Note also that the individual RGB palette + * also endian-specific). Note also that the individual RGB32 palette * components stored in AVFrame.data[1] should be in the range 0..255. * This is important as many custom PAL8 video codecs that were designed * to run on the IBM VGA graphics adapter use 6-bit palette components. @@ -74,6 +77,7 @@ enum AVPixelFormat { #if FF_API_XVMC AV_PIX_FMT_XVMC_MPEG2_MC,///< XVideo Motion Acceleration via common packet passing AV_PIX_FMT_XVMC_MPEG2_IDCT, +#define AV_PIX_FMT_XVMC AV_PIX_FMT_XVMC_MPEG2_IDCT #endif /* FF_API_XVMC */ AV_PIX_FMT_UYVY422, ///< packed YUV 4:2:2, 16bpp, Cb Y0 Cr Y1 AV_PIX_FMT_UYYVYY411, ///< packed YUV 4:1:1, 12bpp, Cb Y0 Y1 Cr Y2 Y3 @@ -108,21 +112,28 @@ enum AVPixelFormat { AV_PIX_FMT_RGB565BE, ///< packed RGB 5:6:5, 16bpp, (msb) 5R 6G 5B(lsb), big-endian AV_PIX_FMT_RGB565LE, ///< packed RGB 5:6:5, 16bpp, (msb) 5R 6G 5B(lsb), little-endian - AV_PIX_FMT_RGB555BE, ///< packed RGB 5:5:5, 16bpp, (msb)1A 5R 5G 5B(lsb), big-endian, most significant bit to 0 - AV_PIX_FMT_RGB555LE, ///< packed RGB 5:5:5, 16bpp, (msb)1A 5R 5G 5B(lsb), little-endian, most significant bit to 0 + AV_PIX_FMT_RGB555BE, ///< packed RGB 5:5:5, 16bpp, (msb)1X 5R 5G 5B(lsb), big-endian , X=unused/undefined + AV_PIX_FMT_RGB555LE, ///< packed RGB 5:5:5, 16bpp, (msb)1X 5R 5G 5B(lsb), little-endian, X=unused/undefined AV_PIX_FMT_BGR565BE, ///< packed BGR 5:6:5, 16bpp, (msb) 5B 6G 5R(lsb), big-endian AV_PIX_FMT_BGR565LE, ///< packed BGR 5:6:5, 16bpp, (msb) 5B 6G 5R(lsb), little-endian - AV_PIX_FMT_BGR555BE, ///< packed BGR 5:5:5, 16bpp, (msb)1A 5B 5G 5R(lsb), big-endian, most significant bit to 1 - AV_PIX_FMT_BGR555LE, ///< packed BGR 5:5:5, 16bpp, (msb)1A 5B 5G 5R(lsb), little-endian, most significant bit to 1 + AV_PIX_FMT_BGR555BE, ///< packed BGR 5:5:5, 16bpp, (msb)1X 5B 5G 5R(lsb), big-endian , X=unused/undefined + AV_PIX_FMT_BGR555LE, ///< packed BGR 5:5:5, 16bpp, (msb)1X 5B 5G 5R(lsb), little-endian, X=unused/undefined #if FF_API_VAAPI + /** @name Deprecated pixel formats */ + /**@{*/ AV_PIX_FMT_VAAPI_MOCO, ///< HW acceleration through VA API at motion compensation entry-point, Picture.data[3] contains a vaapi_render_state struct which contains macroblocks as well as various fields extracted from headers AV_PIX_FMT_VAAPI_IDCT, ///< HW acceleration through VA API at IDCT entry-point, Picture.data[3] contains a vaapi_render_state struct which contains fields extracted from headers AV_PIX_FMT_VAAPI_VLD, ///< HW decoding through VA API, Picture.data[3] contains a VASurfaceID + /**@}*/ AV_PIX_FMT_VAAPI = AV_PIX_FMT_VAAPI_VLD, #else - AV_PIX_FMT_VAAPI, ///< HW decoding through VA API, Picture.data[3] contains a VASurfaceID + /** + * Hardware acceleration through VA-API, data[3] contains a + * VASurfaceID. + */ + AV_PIX_FMT_VAAPI, #endif AV_PIX_FMT_YUV420P16LE, ///< planar YUV 4:2:0, 24bpp, (1 Cr & Cb sample per 2x2 Y samples), little-endian @@ -136,16 +147,23 @@ enum AVPixelFormat { #endif AV_PIX_FMT_DXVA2_VLD, ///< HW decoding through DXVA2, Picture.data[3] contains a LPDIRECT3DSURFACE9 pointer - AV_PIX_FMT_RGB444LE, ///< packed RGB 4:4:4, 16bpp, (msb)4A 4R 4G 4B(lsb), little-endian, most significant bits to 0 - AV_PIX_FMT_RGB444BE, ///< packed RGB 4:4:4, 16bpp, (msb)4A 4R 4G 4B(lsb), big-endian, most significant bits to 0 - AV_PIX_FMT_BGR444LE, ///< packed BGR 4:4:4, 16bpp, (msb)4A 4B 4G 4R(lsb), little-endian, most significant bits to 1 - AV_PIX_FMT_BGR444BE, ///< packed BGR 4:4:4, 16bpp, (msb)4A 4B 4G 4R(lsb), big-endian, most significant bits to 1 + AV_PIX_FMT_RGB444LE, ///< packed RGB 4:4:4, 16bpp, (msb)4X 4R 4G 4B(lsb), little-endian, X=unused/undefined + AV_PIX_FMT_RGB444BE, ///< packed RGB 4:4:4, 16bpp, (msb)4X 4R 4G 4B(lsb), big-endian, X=unused/undefined + AV_PIX_FMT_BGR444LE, ///< packed BGR 4:4:4, 16bpp, (msb)4X 4B 4G 4R(lsb), little-endian, X=unused/undefined + AV_PIX_FMT_BGR444BE, ///< packed BGR 4:4:4, 16bpp, (msb)4X 4B 4G 4R(lsb), big-endian, X=unused/undefined AV_PIX_FMT_YA8, ///< 8bit gray, 8bit alpha AV_PIX_FMT_Y400A = AV_PIX_FMT_YA8, ///< alias for AV_PIX_FMT_YA8 + AV_PIX_FMT_GRAY8A= AV_PIX_FMT_YA8, ///< alias for AV_PIX_FMT_YA8 AV_PIX_FMT_BGR48BE, ///< packed RGB 16:16:16, 48bpp, 16B, 16G, 16R, the 2-byte value for each R/G/B component is stored as big-endian AV_PIX_FMT_BGR48LE, ///< packed RGB 16:16:16, 48bpp, 16B, 16G, 16R, the 2-byte value for each R/G/B component is stored as little-endian + + /** + * The following 12 formats have the disadvantage of needing 1 format for each bit depth. + * Notice that each 9/10 bits sample is stored in 16 bits with extra padding. + * If you want to support multiple bit depths, then using AV_PIX_FMT_YUV420P16* with the bpp stored separately is better. + */ AV_PIX_FMT_YUV420P9BE, ///< planar YUV 4:2:0, 13.5bpp, (1 Cr & Cb sample per 2x2 Y samples), big-endian AV_PIX_FMT_YUV420P9LE, ///< planar YUV 4:2:0, 13.5bpp, (1 Cr & Cb sample per 2x2 Y samples), little-endian AV_PIX_FMT_YUV420P10BE,///< planar YUV 4:2:0, 15bpp, (1 Cr & Cb sample per 2x2 Y samples), big-endian @@ -186,7 +204,9 @@ enum AVPixelFormat { AV_PIX_FMT_YUVA422P16LE, ///< planar YUV 4:2:2 48bpp, (1 Cr & Cb sample per 2x1 Y & A samples, little-endian) AV_PIX_FMT_YUVA444P16BE, ///< planar YUV 4:4:4 64bpp, (1 Cr & Cb sample per 1x1 Y & A samples, big-endian) AV_PIX_FMT_YUVA444P16LE, ///< planar YUV 4:4:4 64bpp, (1 Cr & Cb sample per 1x1 Y & A samples, little-endian) + AV_PIX_FMT_VDPAU, ///< HW acceleration through VDPAU, Picture.data[3] contains a VdpVideoSurface + AV_PIX_FMT_XYZ12LE, ///< packed XYZ 4:4:4, 36 bpp, (msb) 12X, 12Y, 12Z (lsb), the 2-byte value for each X/Y/Z is stored as little-endian, the 4 lower bits are set to 0 AV_PIX_FMT_XYZ12BE, ///< packed XYZ 4:4:4, 36 bpp, (msb) 12X, 12Y, 12Z (lsb), the 2-byte value for each X/Y/Z is stored as big-endian, the 4 lower bits are set to 0 AV_PIX_FMT_NV16, ///< interleaved chroma YUV 4:2:2, 16bpp, (1 Cr & Cb sample per 2x1 Y samples) @@ -227,9 +247,65 @@ enum AVPixelFormat { */ AV_PIX_FMT_CUDA, + AV_PIX_FMT_0RGB=0x123+4,///< packed RGB 8:8:8, 32bpp, XRGBXRGB... X=unused/undefined + AV_PIX_FMT_RGB0, ///< packed RGB 8:8:8, 32bpp, RGBXRGBX... X=unused/undefined + AV_PIX_FMT_0BGR, ///< packed BGR 8:8:8, 32bpp, XBGRXBGR... X=unused/undefined + AV_PIX_FMT_BGR0, ///< packed BGR 8:8:8, 32bpp, BGRXBGRX... X=unused/undefined + + AV_PIX_FMT_YUV420P12BE, ///< planar YUV 4:2:0,18bpp, (1 Cr & Cb sample per 2x2 Y samples), big-endian + AV_PIX_FMT_YUV420P12LE, ///< planar YUV 4:2:0,18bpp, (1 Cr & Cb sample per 2x2 Y samples), little-endian + AV_PIX_FMT_YUV420P14BE, ///< planar YUV 4:2:0,21bpp, (1 Cr & Cb sample per 2x2 Y samples), big-endian + AV_PIX_FMT_YUV420P14LE, ///< planar YUV 4:2:0,21bpp, (1 Cr & Cb sample per 2x2 Y samples), little-endian + AV_PIX_FMT_YUV422P12BE, ///< planar YUV 4:2:2,24bpp, (1 Cr & Cb sample per 2x1 Y samples), big-endian + AV_PIX_FMT_YUV422P12LE, ///< planar YUV 4:2:2,24bpp, (1 Cr & Cb sample per 2x1 Y samples), little-endian + AV_PIX_FMT_YUV422P14BE, ///< planar YUV 4:2:2,28bpp, (1 Cr & Cb sample per 2x1 Y samples), big-endian + AV_PIX_FMT_YUV422P14LE, ///< planar YUV 4:2:2,28bpp, (1 Cr & Cb sample per 2x1 Y samples), little-endian + AV_PIX_FMT_YUV444P12BE, ///< planar YUV 4:4:4,36bpp, (1 Cr & Cb sample per 1x1 Y samples), big-endian + AV_PIX_FMT_YUV444P12LE, ///< planar YUV 4:4:4,36bpp, (1 Cr & Cb sample per 1x1 Y samples), little-endian + AV_PIX_FMT_YUV444P14BE, ///< planar YUV 4:4:4,42bpp, (1 Cr & Cb sample per 1x1 Y samples), big-endian + AV_PIX_FMT_YUV444P14LE, ///< planar YUV 4:4:4,42bpp, (1 Cr & Cb sample per 1x1 Y samples), little-endian + AV_PIX_FMT_GBRP12BE, ///< planar GBR 4:4:4 36bpp, big-endian + AV_PIX_FMT_GBRP12LE, ///< planar GBR 4:4:4 36bpp, little-endian + AV_PIX_FMT_GBRP14BE, ///< planar GBR 4:4:4 42bpp, big-endian + AV_PIX_FMT_GBRP14LE, ///< planar GBR 4:4:4 42bpp, little-endian + AV_PIX_FMT_YUVJ411P, ///< planar YUV 4:1:1, 12bpp, (1 Cr & Cb sample per 4x1 Y samples) full scale (JPEG), deprecated in favor of AV_PIX_FMT_YUV411P and setting color_range + + AV_PIX_FMT_BAYER_BGGR8, ///< bayer, BGBG..(odd line), GRGR..(even line), 8-bit samples */ + AV_PIX_FMT_BAYER_RGGB8, ///< bayer, RGRG..(odd line), GBGB..(even line), 8-bit samples */ + AV_PIX_FMT_BAYER_GBRG8, ///< bayer, GBGB..(odd line), RGRG..(even line), 8-bit samples */ + AV_PIX_FMT_BAYER_GRBG8, ///< bayer, GRGR..(odd line), BGBG..(even line), 8-bit samples */ + AV_PIX_FMT_BAYER_BGGR16LE, ///< bayer, BGBG..(odd line), GRGR..(even line), 16-bit samples, little-endian */ + AV_PIX_FMT_BAYER_BGGR16BE, ///< bayer, BGBG..(odd line), GRGR..(even line), 16-bit samples, big-endian */ + AV_PIX_FMT_BAYER_RGGB16LE, ///< bayer, RGRG..(odd line), GBGB..(even line), 16-bit samples, little-endian */ + AV_PIX_FMT_BAYER_RGGB16BE, ///< bayer, RGRG..(odd line), GBGB..(even line), 16-bit samples, big-endian */ + AV_PIX_FMT_BAYER_GBRG16LE, ///< bayer, GBGB..(odd line), RGRG..(even line), 16-bit samples, little-endian */ + AV_PIX_FMT_BAYER_GBRG16BE, ///< bayer, GBGB..(odd line), RGRG..(even line), 16-bit samples, big-endian */ + AV_PIX_FMT_BAYER_GRBG16LE, ///< bayer, GRGR..(odd line), BGBG..(even line), 16-bit samples, little-endian */ + AV_PIX_FMT_BAYER_GRBG16BE, ///< bayer, GRGR..(odd line), BGBG..(even line), 16-bit samples, big-endian */ +#if !FF_API_XVMC + AV_PIX_FMT_XVMC,///< XVideo Motion Acceleration via common packet passing +#endif /* !FF_API_XVMC */ + AV_PIX_FMT_YUV440P10LE, ///< planar YUV 4:4:0,20bpp, (1 Cr & Cb sample per 1x2 Y samples), little-endian + AV_PIX_FMT_YUV440P10BE, ///< planar YUV 4:4:0,20bpp, (1 Cr & Cb sample per 1x2 Y samples), big-endian + AV_PIX_FMT_YUV440P12LE, ///< planar YUV 4:4:0,24bpp, (1 Cr & Cb sample per 1x2 Y samples), little-endian + AV_PIX_FMT_YUV440P12BE, ///< planar YUV 4:4:0,24bpp, (1 Cr & Cb sample per 1x2 Y samples), big-endian + AV_PIX_FMT_AYUV64LE, ///< packed AYUV 4:4:4,64bpp (1 Cr & Cb sample per 1x1 Y & A samples), little-endian + AV_PIX_FMT_AYUV64BE, ///< packed AYUV 4:4:4,64bpp (1 Cr & Cb sample per 1x1 Y & A samples), big-endian + + AV_PIX_FMT_VIDEOTOOLBOX, ///< hardware decoding through Videotoolbox + + AV_PIX_FMT_P010LE, ///< like NV12, with 10bpp per component, data in the high bits, zeros in the low bits, little-endian + AV_PIX_FMT_P010BE, ///< like NV12, with 10bpp per component, data in the high bits, zeros in the low bits, big-endian + + AV_PIX_FMT_GBRAP12BE, ///< planar GBR 4:4:4:4 48bpp, big-endian + AV_PIX_FMT_GBRAP12LE, ///< planar GBR 4:4:4:4 48bpp, little-endian + AV_PIX_FMT_NB, ///< number of pixel formats, DO NOT USE THIS if you want to link with shared libav* because the number of formats might differ between versions }; +#define AV_PIX_FMT_Y400A AV_PIX_FMT_GRAY8A +#define AV_PIX_FMT_GBR24P AV_PIX_FMT_GBRP + #if AV_HAVE_BIGENDIAN # define AV_PIX_FMT_NE(be, le) AV_PIX_FMT_##be #else @@ -240,6 +316,8 @@ enum AVPixelFormat { #define AV_PIX_FMT_RGB32_1 AV_PIX_FMT_NE(RGBA, ABGR) #define AV_PIX_FMT_BGR32 AV_PIX_FMT_NE(ABGR, RGBA) #define AV_PIX_FMT_BGR32_1 AV_PIX_FMT_NE(BGRA, ARGB) +#define AV_PIX_FMT_0RGB32 AV_PIX_FMT_NE(0RGB, BGR0) +#define AV_PIX_FMT_0BGR32 AV_PIX_FMT_NE(0BGR, RGB0) #define AV_PIX_FMT_GRAY16 AV_PIX_FMT_NE(GRAY16BE, GRAY16LE) #define AV_PIX_FMT_YA16 AV_PIX_FMT_NE(YA16BE, YA16LE) @@ -259,17 +337,33 @@ enum AVPixelFormat { #define AV_PIX_FMT_YUV444P9 AV_PIX_FMT_NE(YUV444P9BE , YUV444P9LE) #define AV_PIX_FMT_YUV420P10 AV_PIX_FMT_NE(YUV420P10BE, YUV420P10LE) #define AV_PIX_FMT_YUV422P10 AV_PIX_FMT_NE(YUV422P10BE, YUV422P10LE) +#define AV_PIX_FMT_YUV440P10 AV_PIX_FMT_NE(YUV440P10BE, YUV440P10LE) #define AV_PIX_FMT_YUV444P10 AV_PIX_FMT_NE(YUV444P10BE, YUV444P10LE) +#define AV_PIX_FMT_YUV420P12 AV_PIX_FMT_NE(YUV420P12BE, YUV420P12LE) +#define AV_PIX_FMT_YUV422P12 AV_PIX_FMT_NE(YUV422P12BE, YUV422P12LE) +#define AV_PIX_FMT_YUV440P12 AV_PIX_FMT_NE(YUV440P12BE, YUV440P12LE) +#define AV_PIX_FMT_YUV444P12 AV_PIX_FMT_NE(YUV444P12BE, YUV444P12LE) +#define AV_PIX_FMT_YUV420P14 AV_PIX_FMT_NE(YUV420P14BE, YUV420P14LE) +#define AV_PIX_FMT_YUV422P14 AV_PIX_FMT_NE(YUV422P14BE, YUV422P14LE) +#define AV_PIX_FMT_YUV444P14 AV_PIX_FMT_NE(YUV444P14BE, YUV444P14LE) #define AV_PIX_FMT_YUV420P16 AV_PIX_FMT_NE(YUV420P16BE, YUV420P16LE) #define AV_PIX_FMT_YUV422P16 AV_PIX_FMT_NE(YUV422P16BE, YUV422P16LE) #define AV_PIX_FMT_YUV444P16 AV_PIX_FMT_NE(YUV444P16BE, YUV444P16LE) #define AV_PIX_FMT_GBRP9 AV_PIX_FMT_NE(GBRP9BE , GBRP9LE) #define AV_PIX_FMT_GBRP10 AV_PIX_FMT_NE(GBRP10BE, GBRP10LE) +#define AV_PIX_FMT_GBRP12 AV_PIX_FMT_NE(GBRP12BE, GBRP12LE) +#define AV_PIX_FMT_GBRP14 AV_PIX_FMT_NE(GBRP14BE, GBRP14LE) #define AV_PIX_FMT_GBRP16 AV_PIX_FMT_NE(GBRP16BE, GBRP16LE) - +#define AV_PIX_FMT_GBRAP12 AV_PIX_FMT_NE(GBRAP12BE, GBRAP12LE) #define AV_PIX_FMT_GBRAP16 AV_PIX_FMT_NE(GBRAP16BE, GBRAP16LE) +#define AV_PIX_FMT_BAYER_BGGR16 AV_PIX_FMT_NE(BAYER_BGGR16BE, BAYER_BGGR16LE) +#define AV_PIX_FMT_BAYER_RGGB16 AV_PIX_FMT_NE(BAYER_RGGB16BE, BAYER_RGGB16LE) +#define AV_PIX_FMT_BAYER_GBRG16 AV_PIX_FMT_NE(BAYER_GBRG16BE, BAYER_GBRG16LE) +#define AV_PIX_FMT_BAYER_GRBG16 AV_PIX_FMT_NE(BAYER_GRBG16BE, BAYER_GRBG16LE) + + #define AV_PIX_FMT_YUVA420P9 AV_PIX_FMT_NE(YUVA420P9BE , YUVA420P9LE) #define AV_PIX_FMT_YUVA422P9 AV_PIX_FMT_NE(YUVA422P9BE , YUVA422P9LE) #define AV_PIX_FMT_YUVA444P9 AV_PIX_FMT_NE(YUVA444P9BE , YUVA444P9LE) @@ -282,24 +376,26 @@ enum AVPixelFormat { #define AV_PIX_FMT_XYZ12 AV_PIX_FMT_NE(XYZ12BE, XYZ12LE) #define AV_PIX_FMT_NV20 AV_PIX_FMT_NE(NV20BE, NV20LE) +#define AV_PIX_FMT_AYUV64 AV_PIX_FMT_NE(AYUV64BE, AYUV64LE) +#define AV_PIX_FMT_P010 AV_PIX_FMT_NE(P010BE, P010LE) /** * Chromaticity coordinates of the source primaries. */ enum AVColorPrimaries { AVCOL_PRI_RESERVED0 = 0, - AVCOL_PRI_BT709 = 1, ///< also ITU-R BT1361 / IEC 61966-2-4 / SMPTE RP177 Annex B + AVCOL_PRI_BT709 = 1, ///< also ITU-R BT1361 / IEC 61966-2-4 / SMPTE RP177 Annex B AVCOL_PRI_UNSPECIFIED = 2, AVCOL_PRI_RESERVED = 3, - AVCOL_PRI_BT470M = 4, ///< also FCC Title 47 Code of Federal Regulations 73.682 (a)(20) - - AVCOL_PRI_BT470BG = 5, ///< also ITU-R BT601-6 625 / ITU-R BT1358 625 / ITU-R BT1700 625 PAL & SECAM - AVCOL_PRI_SMPTE170M = 6, ///< also ITU-R BT601-6 525 / ITU-R BT1358 525 / ITU-R BT1700 NTSC - AVCOL_PRI_SMPTE240M = 7, ///< functionally identical to above - AVCOL_PRI_FILM = 8, ///< colour filters using Illuminant C - AVCOL_PRI_BT2020 = 9, ///< ITU-R BT2020 - AVCOL_PRI_SMPTEST428_1 = 10, ///< SMPTE ST 428-1 (CIE 1931 XYZ) - AVCOL_PRI_NB, ///< Not part of ABI + AVCOL_PRI_BT470M = 4, ///< also FCC Title 47 Code of Federal Regulations 73.682 (a)(20) + + AVCOL_PRI_BT470BG = 5, ///< also ITU-R BT601-6 625 / ITU-R BT1358 625 / ITU-R BT1700 625 PAL & SECAM + AVCOL_PRI_SMPTE170M = 6, ///< also ITU-R BT601-6 525 / ITU-R BT1358 525 / ITU-R BT1700 NTSC + AVCOL_PRI_SMPTE240M = 7, ///< functionally identical to above + AVCOL_PRI_FILM = 8, ///< colour filters using Illuminant C + AVCOL_PRI_BT2020 = 9, ///< ITU-R BT2020 + AVCOL_PRI_SMPTEST428_1= 10, ///< SMPTE ST 428-1 (CIE 1931 XYZ) + AVCOL_PRI_NB, ///< Not part of ABI }; /** @@ -344,6 +440,8 @@ enum AVColorSpace { AVCOL_SPC_BT2020_CL = 10, ///< ITU-R BT2020 constant luminance system AVCOL_SPC_NB, ///< Not part of ABI }; +#define AVCOL_SPC_YCGCO AVCOL_SPC_YCOCG + /** * MPEG vs JPEG YUV range. @@ -358,15 +456,23 @@ enum AVColorRange { /** * Location of chroma samples. * - * X X 3 4 X X are luma samples, - * 1 2 1-6 are possible chroma positions - * X X 5 6 X 0 is undefined/unknown position + * Illustration showing the location of the first (top left) chroma sample of the + * image, the left shows only luma, the right + * shows the location of the chroma sample, the 2 could be imagined to overlay + * each other but are drawn separately due to limitations of ASCII + * + * 1st 2nd 1st 2nd horizontal luma sample positions + * v v v v + * ______ ______ + *1st luma line > |X X ... |3 4 X ... X are luma samples, + * | |1 2 1-6 are possible chroma positions + *2nd luma line > |X X ... |5 6 X ... 0 is undefined/unknown position */ enum AVChromaLocation { AVCHROMA_LOC_UNSPECIFIED = 0, - AVCHROMA_LOC_LEFT = 1, ///< mpeg2/4, h264 default - AVCHROMA_LOC_CENTER = 2, ///< mpeg1, jpeg, h263 - AVCHROMA_LOC_TOPLEFT = 3, ///< DV + AVCHROMA_LOC_LEFT = 1, ///< mpeg2/4 4:2:0, h264 default for 4:2:0 + AVCHROMA_LOC_CENTER = 2, ///< mpeg1 4:2:0, jpeg 4:2:0, h263 4:2:0 + AVCHROMA_LOC_TOPLEFT = 3, ///< ITU-R 601, SMPTE 274M 296M S314M(DV 4:1:1), mpeg2 4:2:2 AVCHROMA_LOC_TOP = 4, AVCHROMA_LOC_BOTTOMLEFT = 5, AVCHROMA_LOC_BOTTOM = 6, diff --git a/libavutil/ppc/cpu.c b/libavutil/ppc/cpu.c index 298185684b..0f1e982624 100644 --- a/libavutil/ppc/cpu.c +++ b/libavutil/ppc/cpu.c @@ -1,28 +1,32 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ +#include "config.h" + #ifdef __APPLE__ #include <sys/sysctl.h> #elif defined(__linux__) #include <asm/cputable.h> #include <linux/auxvec.h> #include <fcntl.h> +#if HAVE_UNISTD_H #include <unistd.h> +#endif #elif defined(__OpenBSD__) #include <sys/param.h> #include <sys/sysctl.h> @@ -33,9 +37,9 @@ #include <proto/exec.h> #endif /* __APPLE__ */ +#include "libavutil/avassert.h" #include "libavutil/cpu.h" #include "libavutil/cpu_internal.h" -#include "config.h" /** * This function MAY rely on signal() or fork() in order to make sure AltiVec @@ -93,6 +97,8 @@ int ff_get_cpu_flags_ppc(void) if (buf[i + 1] & PPC_FEATURE_HAS_POWER8) ret |= AV_CPU_FLAG_POWER8; #endif + if (ret & AV_CPU_FLAG_VSX) + av_assert0(ret & AV_CPU_FLAG_ALTIVEC); goto out; } } @@ -101,7 +107,7 @@ int ff_get_cpu_flags_ppc(void) out: close(fd); return ret; -#elif CONFIG_RUNTIME_CPUDETECT +#elif CONFIG_RUNTIME_CPUDETECT && defined(__linux__) #define PVR_G4_7400 0x000C #define PVR_G5_970 0x0039 #define PVR_G5_970FX 0x003C diff --git a/libavutil/ppc/cpu.h b/libavutil/ppc/cpu.h index a8b823f534..0744157c87 100644 --- a/libavutil/ppc/cpu.h +++ b/libavutil/ppc/cpu.h @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/ppc/float_dsp_altivec.c b/libavutil/ppc/float_dsp_altivec.c index f6918e193f..6aa3e51cef 100644 --- a/libavutil/ppc/float_dsp_altivec.c +++ b/libavutil/ppc/float_dsp_altivec.c @@ -1,20 +1,20 @@ /* * Copyright (c) 2006 Luca Barbato <lu_zero@gentoo.org> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -75,20 +75,18 @@ void ff_vector_fmul_add_altivec(float *dst, const float *src0, int len) { int i; - vec_f d, s0, s1, s2, t0, t1, edges; - vec_u8 align = vec_lvsr(0, dst); - vec_u8 mask = vec_lvsl(0, dst); + vec_f d, ss0, ss1, ss2, t0, t1, edges; for (i = 0; i < len - 3; i += 4) { t0 = vec_ld(0, dst + i); t1 = vec_ld(15, dst + i); - s0 = vec_ld(0, src0 + i); - s1 = vec_ld(0, src1 + i); - s2 = vec_ld(0, src2 + i); - edges = vec_perm(t1, t0, mask); - d = vec_madd(s0, s1, s2); - t1 = vec_perm(d, edges, align); - t0 = vec_perm(edges, d, align); + ss0 = vec_ld(0, src0 + i); + ss1 = vec_ld(0, src1 + i); + ss2 = vec_ld(0, src2 + i); + edges = vec_perm(t1, t0, vcprm(0, 1, 2, 3)); + d = vec_madd(ss0, ss1, ss2); + t1 = vec_perm(d, edges, vcprm(s0,s1,s2,s3)); + t0 = vec_perm(edges, d, vcprm(s0,s1,s2,s3)); vec_st(t1, 15, dst + i); vec_st(t0, 0, dst + i); } diff --git a/libavutil/ppc/float_dsp_altivec.h b/libavutil/ppc/float_dsp_altivec.h index 87d50a8878..e1d530afa5 100644 --- a/libavutil/ppc/float_dsp_altivec.h +++ b/libavutil/ppc/float_dsp_altivec.h @@ -1,20 +1,20 @@ /* * Copyright (c) 2006 Luca Barbato <lu_zero@gentoo.org> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/ppc/float_dsp_init.c b/libavutil/ppc/float_dsp_init.c index 4407962d4d..1d490bd94c 100644 --- a/libavutil/ppc/float_dsp_init.c +++ b/libavutil/ppc/float_dsp_init.c @@ -1,20 +1,20 @@ /* * Copyright (c) 2006 Luca Barbato <lu_zero@gentoo.org> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -28,7 +28,7 @@ av_cold void ff_float_dsp_init_ppc(AVFloatDSPContext *fdsp, int bit_exact) { - if (HAVE_BIGENDIAN && PPC_ALTIVEC(av_get_cpu_flags())) { + if (PPC_ALTIVEC(av_get_cpu_flags())) { fdsp->vector_fmul = ff_vector_fmul_altivec; fdsp->vector_fmul_add = ff_vector_fmul_add_altivec; fdsp->vector_fmul_reverse = ff_vector_fmul_reverse_altivec; @@ -37,13 +37,16 @@ av_cold void ff_float_dsp_init_ppc(AVFloatDSPContext *fdsp, int bit_exact) fdsp->vector_fmul_window = ff_vector_fmul_window_altivec; } } + + // The disabled function below are near identical to altivec and have + // been disabled to reduce code duplication if (PPC_VSX(av_get_cpu_flags())) { - fdsp->vector_fmul = ff_vector_fmul_vsx; +// fdsp->vector_fmul = ff_vector_fmul_vsx; fdsp->vector_fmul_add = ff_vector_fmul_add_vsx; - fdsp->vector_fmul_reverse = ff_vector_fmul_reverse_vsx; +// fdsp->vector_fmul_reverse = ff_vector_fmul_reverse_vsx; - if (!bit_exact) { - fdsp->vector_fmul_window = ff_vector_fmul_window_vsx; - } +// if (!bit_exact) { +// fdsp->vector_fmul_window = ff_vector_fmul_window_vsx; +// } } } diff --git a/libavutil/ppc/float_dsp_vsx.c b/libavutil/ppc/float_dsp_vsx.c index 4135db1391..a91e7c1205 100644 --- a/libavutil/ppc/float_dsp_vsx.c +++ b/libavutil/ppc/float_dsp_vsx.c @@ -1,20 +1,20 @@ /* * Copyright (c) 2015 Luca Barbato <lu_zero@gentoo.org> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/ppc/float_dsp_vsx.h b/libavutil/ppc/float_dsp_vsx.h index 32860dc2c9..5d3cc5ede9 100644 --- a/libavutil/ppc/float_dsp_vsx.h +++ b/libavutil/ppc/float_dsp_vsx.h @@ -1,20 +1,20 @@ /* * Copyright (c) 2015 Luca Barbato <lu_zero@gentoo.org> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/ppc/intreadwrite.h b/libavutil/ppc/intreadwrite.h index 8f8078d2c9..7671676ce9 100644 --- a/libavutil/ppc/intreadwrite.h +++ b/libavutil/ppc/intreadwrite.h @@ -1,20 +1,20 @@ /* * Copyright (c) 2008 Mans Rullgard <mans@mansr.com> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/ppc/timer.h b/libavutil/ppc/timer.h index 8d08d2835b..b28e624566 100644 --- a/libavutil/ppc/timer.h +++ b/libavutil/ppc/timer.h @@ -1,20 +1,20 @@ /* * Copyright (c) 2005 Luca Barbato <lu_zero@gentoo.org> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -25,8 +25,6 @@ #include "config.h" -#if HAVE_INLINE_ASM_LABELS - #define AV_READ_TIME read_time static inline uint64_t read_time(void) @@ -35,12 +33,11 @@ static inline uint64_t read_time(void) /* from section 2.2.1 of the 32-bit PowerPC PEM */ __asm__ volatile( - "1:\n" "mftbu %2\n" "mftb %0\n" "mftbu %1\n" "cmpw %2,%1\n" - "bne 1b\n" + "bne $-0x10\n" : "=r"(tbl), "=r"(tbu), "=r"(temp) : : "cc"); @@ -48,6 +45,4 @@ static inline uint64_t read_time(void) return (((uint64_t)tbu)<<32) | (uint64_t)tbl; } -#endif /* HAVE_INLINE_ASM_LABELS */ - #endif /* AVUTIL_PPC_TIMER_H */ diff --git a/libavutil/ppc/types_altivec.h b/libavutil/ppc/types_altivec.h index 0a4eaf885f..69d8957509 100644 --- a/libavutil/ppc/types_altivec.h +++ b/libavutil/ppc/types_altivec.h @@ -1,20 +1,20 @@ /* * Copyright (c) 2006 Guillaume Poirier <gpoirier@mplayerhq.hu> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/ppc/util_altivec.h b/libavutil/ppc/util_altivec.h index 02cff186f9..5527740010 100644 --- a/libavutil/ppc/util_altivec.h +++ b/libavutil/ppc/util_altivec.h @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -46,9 +46,21 @@ #define WORD_s1 0x14,0x15,0x16,0x17 #define WORD_s2 0x18,0x19,0x1a,0x1b #define WORD_s3 0x1c,0x1d,0x1e,0x1f - #define vcprm(a,b,c,d) (const vec_u8){WORD_ ## a, WORD_ ## b, WORD_ ## c, WORD_ ## d} +#define SWP_W2S0 0x02,0x03,0x00,0x01 +#define SWP_W2S1 0x06,0x07,0x04,0x05 +#define SWP_W2S2 0x0a,0x0b,0x08,0x09 +#define SWP_W2S3 0x0e,0x0f,0x0c,0x0d +#define SWP_W2Ss0 0x12,0x13,0x10,0x11 +#define SWP_W2Ss1 0x16,0x17,0x14,0x15 +#define SWP_W2Ss2 0x1a,0x1b,0x18,0x19 +#define SWP_W2Ss3 0x1e,0x1f,0x1c,0x1d +#define vcswapi2s(a,b,c,d) (const vector unsigned char){SWP_W2S ## a, SWP_W2S ## b, SWP_W2S ## c, SWP_W2S ## d} + +#define vcswapc() \ + (const vector unsigned char){0x0f,0x0e,0x0d,0x0c,0x0b,0x0a,0x09,0x08,0x07,0x06,0x05,0x04,0x03,0x02,0x01,0x00} + // Transpose 8x8 matrix of 16-bit elements (in-place) #define TRANSPOSE8(a,b,c,d,e,f,g,h) \ @@ -85,29 +97,68 @@ do { \ } while (0) +#if HAVE_BIGENDIAN +#define VEC_LD(offset,b) \ + vec_perm(vec_ld(offset, b), vec_ld((offset)+15, b), vec_lvsl(offset, b)) +#else +#define VEC_LD(offset,b) \ + vec_vsx_ld(offset, b) +#endif + /** @brief loads unaligned vector @a *src with offset @a offset and returns it */ -static inline vec_u8 unaligned_load(int offset, uint8_t *src) +#if HAVE_BIGENDIAN +static inline vec_u8 unaligned_load(int offset, const uint8_t *src) { register vec_u8 first = vec_ld(offset, src); register vec_u8 second = vec_ld(offset + 15, src); register vec_u8 mask = vec_lvsl(offset, src); return vec_perm(first, second, mask); } - -/** - * loads vector known misalignment - * @param perm_vec the align permute vector to combine the two loads from lvsl - */ -static inline vec_u8 load_with_perm_vec(int offset, uint8_t *src, vec_u8 perm_vec) +static inline vec_u8 load_with_perm_vec(int offset, const uint8_t *src, vec_u8 perm_vec) { vec_u8 a = vec_ld(offset, src); vec_u8 b = vec_ld(offset + 15, src); return vec_perm(a, b, perm_vec); } +#else +#define unaligned_load(a,b) VEC_LD(a,b) +#define load_with_perm_vec(a,b,c) VEC_LD(a,b) +#endif -#define vec_unaligned_load(b) \ - vec_perm(vec_ld(0, b), vec_ld(15, b), vec_lvsl(0, b)); + +/** + * loads vector known misalignment + * @param perm_vec the align permute vector to combine the two loads from lvsl + */ + +#define vec_unaligned_load(b) VEC_LD(0, b) + +#if HAVE_BIGENDIAN +#define VEC_MERGEH(a, b) vec_mergeh(a, b) +#define VEC_MERGEL(a, b) vec_mergel(a, b) +#else +#define VEC_MERGEH(a, b) vec_mergeh(b, a) +#define VEC_MERGEL(a, b) vec_mergel(b, a) +#endif + +#if HAVE_BIGENDIAN +#define VEC_ST(a,b,c) vec_st(a,b,c) +#else +#define VEC_ST(a,b,c) vec_vsx_st(a,b,c) +#endif + +#if HAVE_BIGENDIAN +#define VEC_SPLAT16(a,b) vec_splat((vec_s16)(a), b) +#else +#define VEC_SPLAT16(a,b) vec_splat((vec_s16)(vec_perm(a, a, vcswapi2s(0,1,2,3))), b) +#endif + +#if HAVE_BIGENDIAN +#define VEC_SLD16(a,b,c) vec_sld(a, b, c) +#else +#define VEC_SLD16(a,b,c) vec_sld(b, a, c) +#endif #endif /* HAVE_ALTIVEC */ diff --git a/libavutil/qsort.h b/libavutil/qsort.h new file mode 100644 index 0000000000..39b7a08852 --- /dev/null +++ b/libavutil/qsort.h @@ -0,0 +1,122 @@ +/* + * copyright (c) 2012 Michael Niedermayer <michaelni@gmx.at> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#ifndef AVUTIL_QSORT_H +#define AVUTIL_QSORT_H + +#include "common.h" + + +/** + * Quicksort + * This sort is fast, and fully inplace but not stable and it is possible + * to construct input that requires O(n^2) time but this is very unlikely to + * happen with non constructed input. + */ +#define AV_QSORT(p, num, type, cmp) do {\ + void *stack[64][2];\ + int sp= 1;\ + stack[0][0] = p;\ + stack[0][1] = (p)+(num)-1;\ + while(sp){\ + type *start= stack[--sp][0];\ + type *end = stack[ sp][1];\ + while(start < end){\ + if(start < end-1) {\ + int checksort=0;\ + type *right = end-2;\ + type *left = start+1;\ + type *mid = start + ((end-start)>>1);\ + if(cmp(start, end) > 0) {\ + if(cmp( end, mid) > 0) FFSWAP(type, *start, *mid);\ + else FFSWAP(type, *start, *end);\ + }else{\ + if(cmp(start, mid) > 0) FFSWAP(type, *start, *mid);\ + else checksort= 1;\ + }\ + if(cmp(mid, end) > 0){ \ + FFSWAP(type, *mid, *end);\ + checksort=0;\ + }\ + if(start == end-2) break;\ + FFSWAP(type, end[-1], *mid);\ + while(left <= right){\ + while(left<=right && cmp(left, end-1) < 0)\ + left++;\ + while(left<=right && cmp(right, end-1) > 0)\ + right--;\ + if(left <= right){\ + FFSWAP(type, *left, *right);\ + left++;\ + right--;\ + }\ + }\ + FFSWAP(type, end[-1], *left);\ + if(checksort && (mid == left-1 || mid == left)){\ + mid= start;\ + while(mid<end && cmp(mid, mid+1) <= 0)\ + mid++;\ + if(mid==end)\ + break;\ + }\ + if(end-left < left-start){\ + stack[sp ][0]= start;\ + stack[sp++][1]= right;\ + start = left+1;\ + }else{\ + stack[sp ][0]= left+1;\ + stack[sp++][1]= end;\ + end = right;\ + }\ + }else{\ + if(cmp(start, end) > 0)\ + FFSWAP(type, *start, *end);\ + break;\ + }\ + }\ + }\ +} while (0) + +/** + * Merge sort, this sort requires a temporary buffer and is stable, its worst + * case time is O(n log n) + * @param p must be a lvalue pointer, this function may exchange it with tmp + * @param tmp must be a lvalue pointer, this function may exchange it with p + */ +#define AV_MSORT(p, tmp, num, type, cmp) do {\ + unsigned i, j, step;\ + for(step=1; step<(num); step+=step){\ + for(i=0; i<(num); i+=2*step){\ + unsigned a[2] = {i, i+step};\ + unsigned end = FFMIN(i+2*step, (num));\ + for(j=i; a[0]<i+step && a[1]<end; j++){\ + int idx= cmp(p+a[0], p+a[1]) > 0;\ + tmp[j] = p[ a[idx]++ ];\ + }\ + if(a[0]>=i+step) a[0] = a[1];\ + for(; j<end; j++){\ + tmp[j] = p[ a[0]++ ];\ + }\ + }\ + FFSWAP(type*, p, tmp);\ + }\ +} while (0) + +#endif /* AVUTIL_QSORT_H */ diff --git a/libavutil/random_seed-test.c b/libavutil/random_seed-test.c new file mode 100644 index 0000000000..182c8a6dd6 --- /dev/null +++ b/libavutil/random_seed-test.c @@ -0,0 +1,46 @@ +/* + * Copyright (c) 2009 Baptiste Coudurier <baptiste.coudurier@gmail.com> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#define TEST 1 +#include "random_seed.c" + +#undef printf +#define N 256 +#include <stdio.h> + +int main(void) +{ + int i, j, retry; + uint32_t seeds[N]; + + for (retry=0; retry<3; retry++){ + for (i=0; i<N; i++){ + seeds[i] = av_get_random_seed(); + for (j=0; j<i; j++) + if (seeds[j] == seeds[i]) + goto retry; + } + printf("seeds OK\n"); + return 0; + retry:; + } + printf("FAIL at %d with %X\n", j, seeds[j]); + return 1; +} diff --git a/libavutil/random_seed.c b/libavutil/random_seed.c index 26884cbcd6..d1ded7b509 100644 --- a/libavutil/random_seed.c +++ b/libavutil/random_seed.c @@ -1,20 +1,20 @@ /* * Copyright (c) 2009 Baptiste Coudurier <baptiste.coudurier@gmail.com> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -23,6 +23,9 @@ #if HAVE_UNISTD_H #include <unistd.h> #endif +#if HAVE_IO_H +#include <io.h> +#endif #if HAVE_CRYPTGENRANDOM #include <windows.h> #include <wincrypt.h> @@ -30,13 +33,18 @@ #include <fcntl.h> #include <math.h> #include <time.h> +#include <string.h> +#include "avassert.h" #include "internal.h" #include "intreadwrite.h" -#include "mem.h" #include "timer.h" #include "random_seed.h" #include "sha.h" +#ifndef TEST +#define TEST 0 +#endif + static int read_random(uint32_t *dst, const char *file) { #if HAVE_UNISTD_H @@ -56,13 +64,26 @@ static int read_random(uint32_t *dst, const char *file) static uint32_t get_generic_seed(void) { - struct AVSHA *sha = av_sha_alloc(); + uint8_t tmp[120]; + struct AVSHA *sha = (void*)tmp; clock_t last_t = 0; static uint64_t i = 0; static uint32_t buffer[512] = { 0 }; unsigned char digest[20]; uint64_t last_i = i; + av_assert0(sizeof(tmp) >= av_sha_size); + + if(TEST){ + memset(buffer, 0, sizeof(buffer)); + last_i = i = 0; + }else{ +#ifdef AV_READ_TIME + buffer[13] ^= AV_READ_TIME(); + buffer[41] ^= AV_READ_TIME()>>32; +#endif + } + for (;;) { clock_t t = clock(); @@ -70,25 +91,23 @@ static uint32_t get_generic_seed(void) buffer[i & 511]++; } else { buffer[++i & 511] += (t - last_t) % 3294638521U; - if (last_i && i - last_i > 4 || i - last_i > 64) + if (last_i && i - last_i > 4 || i - last_i > 64 || TEST && i - last_i > 8) break; } last_t = t; } - if (!sha) { - uint32_t seed = 0; - int j; - // Unable to allocate an sha context, just xor the buffer together - // to create something hopefully unique. - for (j = 0; j < 512; j++) - seed ^= buffer[j]; - return seed; + if(TEST) { + buffer[0] = buffer[1] = 0; + } else { +#ifdef AV_READ_TIME + buffer[111] += AV_READ_TIME(); +#endif } + av_sha_init(sha, 160); - av_sha_update(sha, (const uint8_t *) buffer, sizeof(buffer)); + av_sha_update(sha, (const uint8_t *)buffer, sizeof(buffer)); av_sha_final(sha, digest); - av_free(sha); return AV_RB32(digest) + AV_RB32(digest + 16); } @@ -107,6 +126,10 @@ uint32_t av_get_random_seed(void) } #endif +#if HAVE_ARC4RANDOM + return arc4random(); +#endif + if (read_random(&seed, "/dev/urandom") == sizeof(seed)) return seed; if (read_random(&seed, "/dev/random") == sizeof(seed)) diff --git a/libavutil/random_seed.h b/libavutil/random_seed.h index b1fad13d07..0462a048e0 100644 --- a/libavutil/random_seed.h +++ b/libavutil/random_seed.h @@ -1,20 +1,20 @@ /* * Copyright (c) 2009 Baptiste Coudurier <baptiste.coudurier@gmail.com> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -28,12 +28,11 @@ */ /** - * Get random data. - * - * This function can be called repeatedly to generate more random bits - * as needed. It is generally quite slow, and usually used to seed a - * PRNG. As it uses /dev/urandom and /dev/random, the quality of the - * returned random data depends on the platform. + * Get a seed to use in conjunction with random functions. + * This function tries to provide a good seed at a best effort bases. + * Its possible to call this function multiple times if more bits are needed. + * It can be quite slow, which is why it should only be used as seed for a faster + * PRNG. The quality of the seed depends on the platform. */ uint32_t av_get_random_seed(void); diff --git a/libavutil/rational-test.c b/libavutil/rational-test.c new file mode 100644 index 0000000000..455d9c42b2 --- /dev/null +++ b/libavutil/rational-test.c @@ -0,0 +1,135 @@ +/* + * rational numbers + * Copyright (c) 2003 Michael Niedermayer <michaelni@gmx.at> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "rational.c" + +#include "integer.h" + +int main(void) +{ + AVRational a,b,r; + int i,j,k; + static const int64_t numlist[] = { + INT64_MIN, INT64_MIN+1, INT64_MAX, INT32_MIN, INT32_MAX, 1,0,-1, + 123456789, INT32_MAX-1, INT32_MAX+1LL, UINT32_MAX-1, UINT32_MAX, UINT32_MAX+1LL + }; + + for (a.num = -2; a.num <= 2; a.num++) { + for (a.den = -2; a.den <= 2; a.den++) { + for (b.num = -2; b.num <= 2; b.num++) { + for (b.den = -2; b.den <= 2; b.den++) { + int c = av_cmp_q(a,b); + double d = av_q2d(a) == av_q2d(b) ? + 0 : (av_q2d(a) - av_q2d(b)); + if (d > 0) d = 1; + else if (d < 0) d = -1; + else if (d != d) d = INT_MIN; + if (c != d) + av_log(NULL, AV_LOG_ERROR, "%d/%d %d/%d, %d %f\n", a.num, + a.den, b.num, b.den, c,d); + r = av_sub_q(av_add_q(b,a), b); + if(b.den && (r.num*a.den != a.num*r.den || !r.num != !a.num || !r.den != !a.den)) + av_log(NULL, AV_LOG_ERROR, "%d/%d ", r.num, r.den); + } + } + } + } + + for (i = 0; i < FF_ARRAY_ELEMS(numlist); i++) { + int64_t a = numlist[i]; + + for (j = 0; j < FF_ARRAY_ELEMS(numlist); j++) { + int64_t b = numlist[j]; + if (b<=0) + continue; + for (k = 0; k < FF_ARRAY_ELEMS(numlist); k++) { + int64_t c = numlist[k]; + int64_t res; + AVInteger ai; + + if (c<=0) + continue; + res = av_rescale_rnd(a,b,c, AV_ROUND_ZERO); + + ai = av_mul_i(av_int2i(a), av_int2i(b)); + ai = av_div_i(ai, av_int2i(c)); + + if (av_cmp_i(ai, av_int2i(INT64_MAX)) > 0 && res == INT64_MIN) + continue; + if (av_cmp_i(ai, av_int2i(INT64_MIN)) < 0 && res == INT64_MIN) + continue; + if (av_cmp_i(ai, av_int2i(res)) == 0) + continue; + + // Special exception for INT64_MIN, remove this in case INT64_MIN is handled without off by 1 error + if (av_cmp_i(ai, av_int2i(res-1)) == 0 && a == INT64_MIN) + continue; + + av_log(NULL, AV_LOG_ERROR, "%"PRId64" * %"PRId64" / %"PRId64" = %"PRId64" or %"PRId64"\n", a,b,c, res, av_i2int(ai)); + } + } + } + + for (a.num = 1; a.num <= 10; a.num++) { + for (a.den = 1; a.den <= 10; a.den++) { + if (av_gcd(a.num, a.den) > 1) + continue; + for (b.num = 1; b.num <= 10; b.num++) { + for (b.den = 1; b.den <= 10; b.den++) { + int start; + if (av_gcd(b.num, b.den) > 1) + continue; + if (av_cmp_q(b, a) < 0) + continue; + for (start = 0; start < 10 ; start++) { + int acc= start; + int i; + + for (i = 0; i<100; i++) { + int exact = start + av_rescale_q(i+1, b, a); + acc = av_add_stable(a, acc, b, 1); + if (FFABS(acc - exact) > 2) { + av_log(NULL, AV_LOG_ERROR, "%d/%d %d/%d, %d %d\n", a.num, + a.den, b.num, b.den, acc, exact); + return 1; + } + } + } + } + } + } + } + + for (a.den = 1; a.den < 0x100000000U/3; a.den*=3) { + for (a.num = -1; a.num < (1<<27); a.num += 1 + a.num/100) { + float f = av_int2float(av_q2intfloat(a)); + float f2 = av_q2d(a); + if (fabs(f - f2) > fabs(f)/5000000) { + av_log(NULL, AV_LOG_ERROR, "%d/%d %f %f\n", a.num, + a.den, f, f2); + return 1; + } + + } + } + + return 0; +} diff --git a/libavutil/rational.c b/libavutil/rational.c index 4053936194..35ee08877f 100644 --- a/libavutil/rational.c +++ b/libavutil/rational.c @@ -2,20 +2,20 @@ * rational numbers * Copyright (c) 2003 Michael Niedermayer <michaelni@gmx.at> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -69,6 +69,7 @@ int av_reduce(int *dst_num, int *dst_den, den = next_den; } av_assert2(av_gcd(a1.num, a1.den) <= 1U); + av_assert2(a1.num <= max && a1.den <= max); *dst_num = sign ? -a1.num : a1.num; *dst_den = a1.den; @@ -105,16 +106,20 @@ AVRational av_sub_q(AVRational b, AVRational c) AVRational av_d2q(double d, int max) { AVRational a; -#define LOG2 0.69314718055994530941723212145817656807550013436025 int exponent; int64_t den; if (isnan(d)) return (AVRational) { 0,0 }; - if (isinf(d)) + if (fabs(d) > INT_MAX + 3LL) return (AVRational) { d < 0 ? -1 : 1, 0 }; - exponent = FFMAX( (int)(log(fabs(d) + 1e-20)/LOG2), 0); + frexp(d, &exponent); + exponent = FFMAX(exponent-1, 0); den = 1LL << (61 - exponent); - av_reduce(&a.num, &a.den, (int64_t)(d * den + 0.5), den, max); + // (int64_t)rint() and llrint() do not work with gcc on ia64 and sparc64, + // see Ticket2713 for affected gcc/glibc versions + av_reduce(&a.num, &a.den, floor(d * den + 0.5), den, max); + if ((!a.num || !a.den) && d && max>0 && max<INT_MAX) + av_reduce(&a.num, &a.den, floor(d * den + 0.5), den, INT_MAX); return a; } @@ -143,3 +148,37 @@ int av_find_nearest_q_idx(AVRational q, const AVRational* q_list) return nearest_q_idx; } + +uint32_t av_q2intfloat(AVRational q) { + int64_t n; + int shift; + int sign = 0; + + if (q.den < 0) { + q.den *= -1; + q.num *= -1; + } + if (q.num < 0) { + q.num *= -1; + sign = 1; + } + + if (!q.num && !q.den) return 0xFFC00000; + if (!q.num) return 0; + if (!q.den) return 0x7F800000 | (q.num & 0x80000000); + + shift = 23 + av_log2(q.den) - av_log2(q.num); + if (shift >= 0) n = av_rescale(q.num, 1LL<<shift, q.den); + else n = av_rescale(q.num, 1, ((int64_t)q.den) << -shift); + + shift -= n >= (1<<24); + shift += n < (1<<23); + + if (shift >= 0) n = av_rescale(q.num, 1LL<<shift, q.den); + else n = av_rescale(q.num, 1, ((int64_t)q.den) << -shift); + + av_assert1(n < (1<<24)); + av_assert1(n >= (1<<23)); + + return sign<<31 | (150-shift)<<23 | (n - (1<<23)); +} diff --git a/libavutil/rational.h b/libavutil/rational.h index 5d7dab7fd0..2897469680 100644 --- a/libavutil/rational.h +++ b/libavutil/rational.h @@ -2,20 +2,20 @@ * rational numbers * Copyright (c) 2003 Michael Niedermayer <michaelni@gmx.at> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -46,6 +46,17 @@ typedef struct AVRational{ } AVRational; /** + * Create a rational. + * Useful for compilers that do not support compound literals. + * @note The return value is not reduced. + */ +static inline AVRational av_make_q(int num, int den) +{ + AVRational r = { num, den }; + return r; +} + +/** * Compare two rationals. * @param a first rational * @param b second rational @@ -55,7 +66,7 @@ typedef struct AVRational{ static inline int av_cmp_q(AVRational a, AVRational b){ const int64_t tmp= a.num * (int64_t)b.den - b.num * (int64_t)a.den; - if(tmp) return ((tmp ^ a.den ^ b.den)>>63)|1; + if(tmp) return (int)((tmp ^ a.den ^ b.den)>>63)|1; else if(b.den && a.den) return 0; else if(a.num && b.num) return (a.num>>31) - (b.num>>31); else return INT_MIN; @@ -149,6 +160,13 @@ int av_nearer_q(AVRational q, AVRational q1, AVRational q2); int av_find_nearest_q_idx(AVRational q, const AVRational* q_list); /** + * Converts a AVRational to a IEEE 32bit float. + * + * The float is returned in a uint32_t and its value is platform indepenant. + */ +uint32_t av_q2intfloat(AVRational q); + +/** * @} */ diff --git a/libavutil/rc4.c b/libavutil/rc4.c index 36b0de9f53..ffcb112142 100644 --- a/libavutil/rc4.c +++ b/libavutil/rc4.c @@ -4,20 +4,20 @@ * * loosely based on LibTomCrypt by Tom St Denis * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ #include "avutil.h" @@ -25,13 +25,6 @@ #include "mem.h" #include "rc4.h" -#if !FF_API_CRYPTO_CONTEXT -struct AVRC4 { - uint8_t state[256]; - int x, y; -}; -#endif - AVRC4 *av_rc4_alloc(void) { return av_mallocz(sizeof(struct AVRC4)); @@ -43,7 +36,7 @@ int av_rc4_init(AVRC4 *r, const uint8_t *key, int key_bits, int decrypt) { uint8_t *state = r->state; int keylen = key_bits >> 3; if (key_bits & 7) - return -1; + return AVERROR(EINVAL); for (i = 0; i < 256; i++) state[i] = i; y = 0; diff --git a/libavutil/rc4.h b/libavutil/rc4.h index 38ae52f298..029cd2ad58 100644 --- a/libavutil/rc4.h +++ b/libavutil/rc4.h @@ -1,20 +1,20 @@ /* * RC4 encryption/decryption/pseudo-random number generator * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -22,7 +22,6 @@ #define AVUTIL_RC4_H #include <stdint.h> -#include "version.h" /** * @defgroup lavu_rc4 RC4 @@ -30,14 +29,10 @@ * @{ */ -#if FF_API_CRYPTO_CONTEXT typedef struct AVRC4 { uint8_t state[256]; int x, y; } AVRC4; -#else -typedef struct AVRC4 AVRC4; -#endif /** * Allocate an AVRC4 context. diff --git a/libavutil/replaygain.h b/libavutil/replaygain.h index 15da873770..b49bf1a3d9 100644 --- a/libavutil/replaygain.h +++ b/libavutil/replaygain.h @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/reverse.c b/libavutil/reverse.c new file mode 100644 index 0000000000..105eb03dda --- /dev/null +++ b/libavutil/reverse.c @@ -0,0 +1,40 @@ +/* + * Copyright (c) 2002-2004 Michael Niedermayer <michaelni@gmx.at> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include <stdint.h> + +const uint8_t ff_reverse[256] = { +0x00,0x80,0x40,0xC0,0x20,0xA0,0x60,0xE0,0x10,0x90,0x50,0xD0,0x30,0xB0,0x70,0xF0, +0x08,0x88,0x48,0xC8,0x28,0xA8,0x68,0xE8,0x18,0x98,0x58,0xD8,0x38,0xB8,0x78,0xF8, +0x04,0x84,0x44,0xC4,0x24,0xA4,0x64,0xE4,0x14,0x94,0x54,0xD4,0x34,0xB4,0x74,0xF4, +0x0C,0x8C,0x4C,0xCC,0x2C,0xAC,0x6C,0xEC,0x1C,0x9C,0x5C,0xDC,0x3C,0xBC,0x7C,0xFC, +0x02,0x82,0x42,0xC2,0x22,0xA2,0x62,0xE2,0x12,0x92,0x52,0xD2,0x32,0xB2,0x72,0xF2, +0x0A,0x8A,0x4A,0xCA,0x2A,0xAA,0x6A,0xEA,0x1A,0x9A,0x5A,0xDA,0x3A,0xBA,0x7A,0xFA, +0x06,0x86,0x46,0xC6,0x26,0xA6,0x66,0xE6,0x16,0x96,0x56,0xD6,0x36,0xB6,0x76,0xF6, +0x0E,0x8E,0x4E,0xCE,0x2E,0xAE,0x6E,0xEE,0x1E,0x9E,0x5E,0xDE,0x3E,0xBE,0x7E,0xFE, +0x01,0x81,0x41,0xC1,0x21,0xA1,0x61,0xE1,0x11,0x91,0x51,0xD1,0x31,0xB1,0x71,0xF1, +0x09,0x89,0x49,0xC9,0x29,0xA9,0x69,0xE9,0x19,0x99,0x59,0xD9,0x39,0xB9,0x79,0xF9, +0x05,0x85,0x45,0xC5,0x25,0xA5,0x65,0xE5,0x15,0x95,0x55,0xD5,0x35,0xB5,0x75,0xF5, +0x0D,0x8D,0x4D,0xCD,0x2D,0xAD,0x6D,0xED,0x1D,0x9D,0x5D,0xDD,0x3D,0xBD,0x7D,0xFD, +0x03,0x83,0x43,0xC3,0x23,0xA3,0x63,0xE3,0x13,0x93,0x53,0xD3,0x33,0xB3,0x73,0xF3, +0x0B,0x8B,0x4B,0xCB,0x2B,0xAB,0x6B,0xEB,0x1B,0x9B,0x5B,0xDB,0x3B,0xBB,0x7B,0xFB, +0x07,0x87,0x47,0xC7,0x27,0xA7,0x67,0xE7,0x17,0x97,0x57,0xD7,0x37,0xB7,0x77,0xF7, +0x0F,0x8F,0x4F,0xCF,0x2F,0xAF,0x6F,0xEF,0x1F,0x9F,0x5F,0xDF,0x3F,0xBF,0x7F,0xFF, +}; diff --git a/libavutil/ripemd-test.c b/libavutil/ripemd-test.c new file mode 100644 index 0000000000..5f1c39cec4 --- /dev/null +++ b/libavutil/ripemd-test.c @@ -0,0 +1,74 @@ +/* + * Copyright (C) 2007 Michael Niedermayer <michaelni@gmx.at> + * Copyright (C) 2013 James Almer + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "ripemd.c" + +#include <stdio.h> + +int main(void) +{ + int i, j, k; + AVRIPEMD ctx; + unsigned char digest[40]; + static const int lengths[4] = { 128, 160, 256, 320 }; + + for (j = 0; j < 4; j++) { + printf("Testing RIPEMD-%d\n", lengths[j]); + for (k = 0; k < 3; k++) { + av_ripemd_init(&ctx, lengths[j]); + if (k == 0) + av_ripemd_update(&ctx, "abc", 3); + else if (k == 1) + av_ripemd_update(&ctx, "abcdbcdecdefdefgefghfghighijhijkijkljklmklmnlmnomnopnopq", 56); + else + for (i = 0; i < 1000*1000; i++) + av_ripemd_update(&ctx, "a", 1); + av_ripemd_final(&ctx, digest); + for (i = 0; i < lengths[j] >> 3; i++) + printf("%02X", digest[i]); + putchar('\n'); + } + switch (j) { //test vectors (from ISO:IEC 10118-3 (2004) and http://homes.esat.kuleuven.be/~bosselae/ripemd160.html) + case 0: + printf("c14a1219 9c66e4ba 84636b0f 69144c77\n" + "a1aa0689 d0fafa2d dc22e88b 49133a06\n" + "4a7f5723 f954eba1 216c9d8f 6320431f\n"); + break; + case 1: + printf("8eb208f7 e05d987a 9b044a8e 98c6b087 f15a0bfc\n" + "12a05338 4a9c0c88 e405a06c 27dcf49a da62eb2b\n" + "52783243 c1697bdb e16d37f9 7f68f083 25dc1528\n"); + break; + case 2: + printf("afbd6e22 8b9d8cbb cef5ca2d 03e6dba1 0ac0bc7d cbe4680e 1e42d2e9 75459b65\n" + "38430455 83aac6c8 c8d91285 73e7a980 9afb2a0f 34ccc36e a9e72f16 f6368e3f\n" + "ac953744 e10e3151 4c150d4d 8d7b6773 42e33399 788296e4 3ae4850c e4f97978\n"); + break; + case 3: + printf("de4c01b3 054f8930 a79d09ae 738e9230 1e5a1708 5beffdc1 b8d11671 3e74f82f a942d64c dbc4682d\n" + "d034a795 0cf72202 1ba4b84d f769a5de 2060e259 df4c9bb4 a4268c0e 935bbc74 70a969c9 d072a1ac\n" + "bdee37f4 371e2064 6b8b0d86 2dda1629 2ae36f40 965e8c85 09e63d1d bddecc50 3e2b63eb 9245bb66\n"); + break; + } + } + + return 0; +} diff --git a/libavutil/ripemd.c b/libavutil/ripemd.c new file mode 100644 index 0000000000..b0db2970db --- /dev/null +++ b/libavutil/ripemd.c @@ -0,0 +1,551 @@ +/* + * Copyright (C) 2007 Michael Niedermayer <michaelni@gmx.at> + * Copyright (C) 2013 James Almer + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include <string.h> + +#include "attributes.h" +#include "avutil.h" +#include "bswap.h" +#include "intreadwrite.h" +#include "ripemd.h" +#include "mem.h" + +/** hash context */ +typedef struct AVRIPEMD { + uint8_t digest_len; ///< digest length in 32-bit words + uint64_t count; ///< number of bytes in buffer + uint8_t buffer[64]; ///< 512-bit buffer of input values used in hash updating + uint32_t state[10]; ///< current hash value + /** function used to update hash for 512-bit input block */ + void (*transform)(uint32_t *state, const uint8_t buffer[64]); +} AVRIPEMD; + +const int av_ripemd_size = sizeof(AVRIPEMD); + +struct AVRIPEMD *av_ripemd_alloc(void) +{ + return av_mallocz(sizeof(struct AVRIPEMD)); +} + +static const uint32_t KA[4] = { + 0x5a827999, 0x6ed9eba1, 0x8f1bbcdc, 0xa953fd4e +}; + +static const uint32_t KB[4] = { + 0x50a28be6, 0x5c4dd124, 0x6d703ef3, 0x7a6d76e9 +}; + +static const int ROTA[80] = { + 11, 14, 15, 12, 5, 8, 7 , 9, 11, 13, 14, 15, 6, 7, 9, 8, + 7 , 6, 8, 13, 11, 9, 7, 15, 7, 12, 15, 9, 11, 7, 13, 12, + 11, 13, 6, 7, 14, 9, 13, 15, 14, 8, 13, 6, 5, 12, 7, 5, + 11, 12, 14, 15, 14, 15, 9, 8, 9, 14, 5, 6, 8, 6, 5, 12, + 9, 15, 5, 11, 6, 8, 13, 12, 5, 12, 13, 14, 11, 8, 5, 6 +}; + +static const int ROTB[80] = { + 8, 9, 9, 11, 13, 15, 15, 5, 7, 7, 8, 11, 14, 14, 12, 6, + 9, 13, 15, 7, 12, 8, 9, 11, 7, 7, 12, 7, 6, 15, 13, 11, + 9, 7, 15, 11, 8, 6, 6, 14, 12, 13, 5, 14, 13, 13, 7, 5, + 15, 5, 8, 11, 14, 14, 6, 14, 6, 9, 12, 9, 12, 5, 15, 8, + 8, 5, 12, 9, 12, 5, 14, 6, 8, 13, 6, 5, 15, 13, 11, 11 +}; + +static const int WA[80] = { + 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, + 7, 4, 13, 1, 10, 6, 15, 3, 12, 0, 9, 5, 2, 14, 11, 8, + 3, 10, 14, 4, 9, 15, 8, 1, 2, 7, 0, 6, 13, 11, 5, 12, + 1, 9, 11, 10, 0, 8, 12, 4, 13, 3, 7, 15, 14, 5, 6, 2, + 4, 0, 5, 9, 7, 12, 2, 10, 14, 1, 3, 8, 11, 6, 15, 13 +}; + +static const int WB[80] = { + 5, 14, 7, 0, 9, 2, 11, 4, 13, 6, 15, 8, 1, 10, 3, 12, + 6, 11, 3, 7, 0, 13, 5, 10, 14, 15, 8, 12, 4, 9, 1, 2, + 15, 5, 1, 3, 7, 14, 6, 9, 11, 8, 12, 2, 10, 0, 4, 13, + 8, 6, 4, 1, 3, 11, 15, 0, 5, 12, 2, 13, 9, 7, 10, 14, + 12, 15, 10, 4, 1, 5, 8, 7, 6, 2, 13, 14, 0, 3, 9, 11 +}; + +#define rol(value, bits) (((value) << (bits)) | ((value) >> (32 - (bits)))) + +#define ROUND128_0_TO_15(a,b,c,d,e,f,g,h) \ + a = rol(a + (( b ^ c ^ d) + block[WA[n]]), ROTA[n]); \ + e = rol(e + ((((f ^ g) & h) ^ g) + block[WB[n]] + KB[0]), ROTB[n]); \ + n++ + +#define ROUND128_16_TO_31(a,b,c,d,e,f,g,h) \ + a = rol(a + ((((c ^ d) & b) ^ d) + block[WA[n]] + KA[0]), ROTA[n]); \ + e = rol(e + (((~g | f) ^ h) + block[WB[n]] + KB[1]), ROTB[n]); \ + n++ + +#define ROUND128_32_TO_47(a,b,c,d,e,f,g,h) \ + a = rol(a + (((~c | b) ^ d) + block[WA[n]] + KA[1]), ROTA[n]); \ + e = rol(e + ((((g ^ h) & f) ^ h) + block[WB[n]] + KB[2]), ROTB[n]); \ + n++ + +#define ROUND128_48_TO_63(a,b,c,d,e,f,g,h) \ + a = rol(a + ((((b ^ c) & d) ^ c) + block[WA[n]] + KA[2]), ROTA[n]); \ + e = rol(e + (( f ^ g ^ h) + block[WB[n]]), ROTB[n]); \ + n++ + +#define R128_0 \ + ROUND128_0_TO_15(a,b,c,d,e,f,g,h); \ + ROUND128_0_TO_15(d,a,b,c,h,e,f,g); \ + ROUND128_0_TO_15(c,d,a,b,g,h,e,f); \ + ROUND128_0_TO_15(b,c,d,a,f,g,h,e) + +#define R128_16 \ + ROUND128_16_TO_31(a,b,c,d,e,f,g,h); \ + ROUND128_16_TO_31(d,a,b,c,h,e,f,g); \ + ROUND128_16_TO_31(c,d,a,b,g,h,e,f); \ + ROUND128_16_TO_31(b,c,d,a,f,g,h,e) + +#define R128_32 \ + ROUND128_32_TO_47(a,b,c,d,e,f,g,h); \ + ROUND128_32_TO_47(d,a,b,c,h,e,f,g); \ + ROUND128_32_TO_47(c,d,a,b,g,h,e,f); \ + ROUND128_32_TO_47(b,c,d,a,f,g,h,e) + +#define R128_48 \ + ROUND128_48_TO_63(a,b,c,d,e,f,g,h); \ + ROUND128_48_TO_63(d,a,b,c,h,e,f,g); \ + ROUND128_48_TO_63(c,d,a,b,g,h,e,f); \ + ROUND128_48_TO_63(b,c,d,a,f,g,h,e) + +static void ripemd128_transform(uint32_t *state, const uint8_t buffer[64]) +{ + uint32_t a, b, c, d, e, f, g, h, av_unused t; + uint32_t block[16]; + int n; + + a = e = state[0]; + b = f = state[1]; + c = g = state[2]; + d = h = state[3]; + + for (n = 0; n < 16; n++) + block[n] = AV_RL32(buffer + 4 * n); + n = 0; + +#if CONFIG_SMALL + for (; n < 16;) { + ROUND128_0_TO_15(a,b,c,d,e,f,g,h); + t = d; d = c; c = b; b = a; a = t; + t = h; h = g; g = f; f = e; e = t; + } + + for (; n < 32;) { + ROUND128_16_TO_31(a,b,c,d,e,f,g,h); + t = d; d = c; c = b; b = a; a = t; + t = h; h = g; g = f; f = e; e = t; + } + + for (; n < 48;) { + ROUND128_32_TO_47(a,b,c,d,e,f,g,h); + t = d; d = c; c = b; b = a; a = t; + t = h; h = g; g = f; f = e; e = t; + } + + for (; n < 64;) { + ROUND128_48_TO_63(a,b,c,d,e,f,g,h); + t = d; d = c; c = b; b = a; a = t; + t = h; h = g; g = f; f = e; e = t; + } +#else + + R128_0; R128_0; R128_0; R128_0; + + R128_16; R128_16; R128_16; R128_16; + + R128_32; R128_32; R128_32; R128_32; + + R128_48; R128_48; R128_48; R128_48; +#endif + + h += c + state[1]; + state[1] = state[2] + d + e; + state[2] = state[3] + a + f; + state[3] = state[0] + b + g; + state[0] = h; +} + +static void ripemd256_transform(uint32_t *state, const uint8_t buffer[64]) +{ + uint32_t a, b, c, d, e, f, g, h, av_unused t; + uint32_t block[16]; + int n; + + a = state[0]; b = state[1]; c = state[2]; d = state[3]; + e = state[4]; f = state[5]; g = state[6]; h = state[7]; + + for (n = 0; n < 16; n++) + block[n] = AV_RL32(buffer + 4 * n); + n = 0; + +#if CONFIG_SMALL + for (; n < 16;) { + ROUND128_0_TO_15(a,b,c,d,e,f,g,h); + t = d; d = c; c = b; b = a; a = t; + t = h; h = g; g = f; f = e; e = t; + } + FFSWAP(uint32_t, a, e); + + for (; n < 32;) { + ROUND128_16_TO_31(a,b,c,d,e,f,g,h); + t = d; d = c; c = b; b = a; a = t; + t = h; h = g; g = f; f = e; e = t; + } + FFSWAP(uint32_t, b, f); + + for (; n < 48;) { + ROUND128_32_TO_47(a,b,c,d,e,f,g,h); + t = d; d = c; c = b; b = a; a = t; + t = h; h = g; g = f; f = e; e = t; + } + FFSWAP(uint32_t, c, g); + + for (; n < 64;) { + ROUND128_48_TO_63(a,b,c,d,e,f,g,h); + t = d; d = c; c = b; b = a; a = t; + t = h; h = g; g = f; f = e; e = t; + } + FFSWAP(uint32_t, d, h); +#else + + R128_0; R128_0; R128_0; R128_0; + FFSWAP(uint32_t, a, e); + + R128_16; R128_16; R128_16; R128_16; + FFSWAP(uint32_t, b, f); + + R128_32; R128_32; R128_32; R128_32; + FFSWAP(uint32_t, c, g); + + R128_48; R128_48; R128_48; R128_48; + FFSWAP(uint32_t, d, h); +#endif + + state[0] += a; state[1] += b; state[2] += c; state[3] += d; + state[4] += e; state[5] += f; state[6] += g; state[7] += h; +} + +#define ROTATE(x,y) \ + x = rol(x, 10); \ + y = rol(y, 10); \ + n++ + +#define ROUND160_0_TO_15(a,b,c,d,e,f,g,h,i,j) \ + a = rol(a + (( b ^ c ^ d) + block[WA[n]]), ROTA[n]) + e; \ + f = rol(f + (((~i | h) ^ g) + block[WB[n]] + KB[0]), ROTB[n]) + j; \ + ROTATE(c,h) + +#define ROUND160_16_TO_31(a,b,c,d,e,f,g,h,i,j) \ + a = rol(a + ((((c ^ d) & b) ^ d) + block[WA[n]] + KA[0]), ROTA[n]) + e; \ + f = rol(f + ((((g ^ h) & i) ^ h) + block[WB[n]] + KB[1]), ROTB[n]) + j; \ + ROTATE(c,h) + +#define ROUND160_32_TO_47(a,b,c,d,e,f,g,h,i,j) \ + a = rol(a + (((~c | b) ^ d) + block[WA[n]] + KA[1]), ROTA[n]) + e; \ + f = rol(f + (((~h | g) ^ i) + block[WB[n]] + KB[2]), ROTB[n]) + j; \ + ROTATE(c,h) + +#define ROUND160_48_TO_63(a,b,c,d,e,f,g,h,i,j) \ + a = rol(a + ((((b ^ c) & d) ^ c) + block[WA[n]] + KA[2]), ROTA[n]) + e; \ + f = rol(f + ((((h ^ i) & g) ^ i) + block[WB[n]] + KB[3]), ROTB[n]) + j; \ + ROTATE(c,h) + +#define ROUND160_64_TO_79(a,b,c,d,e,f,g,h,i,j) \ + a = rol(a + (((~d | c) ^ b) + block[WA[n]] + KA[3]), ROTA[n]) + e; \ + f = rol(f + (( g ^ h ^ i) + block[WB[n]]), ROTB[n]) + j; \ + ROTATE(c,h) + +#define R160_0 \ + ROUND160_0_TO_15(a,b,c,d,e,f,g,h,i,j); \ + ROUND160_0_TO_15(e,a,b,c,d,j,f,g,h,i); \ + ROUND160_0_TO_15(d,e,a,b,c,i,j,f,g,h); \ + ROUND160_0_TO_15(c,d,e,a,b,h,i,j,f,g); \ + ROUND160_0_TO_15(b,c,d,e,a,g,h,i,j,f) + +#define R160_16 \ + ROUND160_16_TO_31(e,a,b,c,d,j,f,g,h,i); \ + ROUND160_16_TO_31(d,e,a,b,c,i,j,f,g,h); \ + ROUND160_16_TO_31(c,d,e,a,b,h,i,j,f,g); \ + ROUND160_16_TO_31(b,c,d,e,a,g,h,i,j,f); \ + ROUND160_16_TO_31(a,b,c,d,e,f,g,h,i,j) + +#define R160_32 \ + ROUND160_32_TO_47(d,e,a,b,c,i,j,f,g,h); \ + ROUND160_32_TO_47(c,d,e,a,b,h,i,j,f,g); \ + ROUND160_32_TO_47(b,c,d,e,a,g,h,i,j,f); \ + ROUND160_32_TO_47(a,b,c,d,e,f,g,h,i,j); \ + ROUND160_32_TO_47(e,a,b,c,d,j,f,g,h,i) + +#define R160_48 \ + ROUND160_48_TO_63(c,d,e,a,b,h,i,j,f,g); \ + ROUND160_48_TO_63(b,c,d,e,a,g,h,i,j,f); \ + ROUND160_48_TO_63(a,b,c,d,e,f,g,h,i,j); \ + ROUND160_48_TO_63(e,a,b,c,d,j,f,g,h,i); \ + ROUND160_48_TO_63(d,e,a,b,c,i,j,f,g,h) + +#define R160_64 \ + ROUND160_64_TO_79(b,c,d,e,a,g,h,i,j,f); \ + ROUND160_64_TO_79(a,b,c,d,e,f,g,h,i,j); \ + ROUND160_64_TO_79(e,a,b,c,d,j,f,g,h,i); \ + ROUND160_64_TO_79(d,e,a,b,c,i,j,f,g,h); \ + ROUND160_64_TO_79(c,d,e,a,b,h,i,j,f,g) + +static void ripemd160_transform(uint32_t *state, const uint8_t buffer[64]) +{ + uint32_t a, b, c, d, e, f, g, h, i, j, av_unused t; + uint32_t block[16]; + int n; + + a = f = state[0]; + b = g = state[1]; + c = h = state[2]; + d = i = state[3]; + e = j = state[4]; + + for (n = 0; n < 16; n++) + block[n] = AV_RL32(buffer + 4 * n); + n = 0; + +#if CONFIG_SMALL + for (; n < 16;) { + ROUND160_0_TO_15(a,b,c,d,e,f,g,h,i,j); + t = e; e = d; d = c; c = b; b = a; a = t; + t = j; j = i; i = h; h = g; g = f; f = t; + } + + for (; n < 32;) { + ROUND160_16_TO_31(a,b,c,d,e,f,g,h,i,j); + t = e; e = d; d = c; c = b; b = a; a = t; + t = j; j = i; i = h; h = g; g = f; f = t; + } + + for (; n < 48;) { + ROUND160_32_TO_47(a,b,c,d,e,f,g,h,i,j); + t = e; e = d; d = c; c = b; b = a; a = t; + t = j; j = i; i = h; h = g; g = f; f = t; + } + + for (; n < 64;) { + ROUND160_48_TO_63(a,b,c,d,e,f,g,h,i,j); + t = e; e = d; d = c; c = b; b = a; a = t; + t = j; j = i; i = h; h = g; g = f; f = t; + } + + for (; n < 80;) { + ROUND160_64_TO_79(a,b,c,d,e,f,g,h,i,j); + t = e; e = d; d = c; c = b; b = a; a = t; + t = j; j = i; i = h; h = g; g = f; f = t; + } +#else + + R160_0; R160_0; R160_0; + ROUND160_0_TO_15(a,b,c,d,e,f,g,h,i,j); + + R160_16; R160_16; R160_16; + ROUND160_16_TO_31(e,a,b,c,d,j,f,g,h,i); + + R160_32; R160_32; R160_32; + ROUND160_32_TO_47(d,e,a,b,c,i,j,f,g,h); + + R160_48; R160_48; R160_48; + ROUND160_48_TO_63(c,d,e,a,b,h,i,j,f,g); + + R160_64; R160_64; R160_64; + ROUND160_64_TO_79(b,c,d,e,a,g,h,i,j,f); +#endif + + i += c + state[1]; + state[1] = state[2] + d + j; + state[2] = state[3] + e + f; + state[3] = state[4] + a + g; + state[4] = state[0] + b + h; + state[0] = i; +} + +static void ripemd320_transform(uint32_t *state, const uint8_t buffer[64]) +{ + uint32_t a, b, c, d, e, f, g, h, i, j, av_unused t; + uint32_t block[16]; + int n; + + a = state[0]; b = state[1]; c = state[2]; d = state[3]; e = state[4]; + f = state[5]; g = state[6]; h = state[7]; i = state[8]; j = state[9]; + + for (n = 0; n < 16; n++) + block[n] = AV_RL32(buffer + 4 * n); + n = 0; + +#if CONFIG_SMALL + for (; n < 16;) { + ROUND160_0_TO_15(a,b,c,d,e,f,g,h,i,j); + t = e; e = d; d = c; c = b; b = a; a = t; + t = j; j = i; i = h; h = g; g = f; f = t; + } + FFSWAP(uint32_t, b, g); + + for (; n < 32;) { + ROUND160_16_TO_31(a,b,c,d,e,f,g,h,i,j); + t = e; e = d; d = c; c = b; b = a; a = t; + t = j; j = i; i = h; h = g; g = f; f = t; + } + FFSWAP(uint32_t, d, i); + + for (; n < 48;) { + ROUND160_32_TO_47(a,b,c,d,e,f,g,h,i,j); + t = e; e = d; d = c; c = b; b = a; a = t; + t = j; j = i; i = h; h = g; g = f; f = t; + } + FFSWAP(uint32_t, a, f); + + for (; n < 64;) { + ROUND160_48_TO_63(a,b,c,d,e,f,g,h,i,j); + t = e; e = d; d = c; c = b; b = a; a = t; + t = j; j = i; i = h; h = g; g = f; f = t; + } + FFSWAP(uint32_t, c, h); + + for (; n < 80;) { + ROUND160_64_TO_79(a,b,c,d,e,f,g,h,i,j); + t = e; e = d; d = c; c = b; b = a; a = t; + t = j; j = i; i = h; h = g; g = f; f = t; + } + FFSWAP(uint32_t, e, j); +#else + + R160_0; R160_0; R160_0; + ROUND160_0_TO_15(a,b,c,d,e,f,g,h,i,j); + FFSWAP(uint32_t, a, f); + + R160_16; R160_16; R160_16; + ROUND160_16_TO_31(e,a,b,c,d,j,f,g,h,i); + FFSWAP(uint32_t, b, g); + + R160_32; R160_32; R160_32; + ROUND160_32_TO_47(d,e,a,b,c,i,j,f,g,h); + FFSWAP(uint32_t, c, h); + + R160_48; R160_48; R160_48; + ROUND160_48_TO_63(c,d,e,a,b,h,i,j,f,g); + FFSWAP(uint32_t, d, i); + + R160_64; R160_64; R160_64; + ROUND160_64_TO_79(b,c,d,e,a,g,h,i,j,f); + FFSWAP(uint32_t, e, j); +#endif + + state[0] += a; state[1] += b; state[2] += c; state[3] += d; state[4] += e; + state[5] += f; state[6] += g; state[7] += h; state[8] += i; state[9] += j; +} + +av_cold int av_ripemd_init(AVRIPEMD *ctx, int bits) +{ + ctx->digest_len = bits >> 5; + switch (bits) { + case 128: // RIPEMD-128 + ctx->state[0] = 0x67452301; + ctx->state[1] = 0xEFCDAB89; + ctx->state[2] = 0x98BADCFE; + ctx->state[3] = 0x10325476; + ctx->transform = ripemd128_transform; + break; + case 160: // RIPEMD-160 + ctx->state[0] = 0x67452301; + ctx->state[1] = 0xEFCDAB89; + ctx->state[2] = 0x98BADCFE; + ctx->state[3] = 0x10325476; + ctx->state[4] = 0xC3D2E1F0; + ctx->transform = ripemd160_transform; + break; + case 256: // RIPEMD-256 + ctx->state[0] = 0x67452301; + ctx->state[1] = 0xEFCDAB89; + ctx->state[2] = 0x98BADCFE; + ctx->state[3] = 0x10325476; + ctx->state[4] = 0x76543210; + ctx->state[5] = 0xFEDCBA98; + ctx->state[6] = 0x89ABCDEF; + ctx->state[7] = 0x01234567; + ctx->transform = ripemd256_transform; + break; + case 320: // RIPEMD-320 + ctx->state[0] = 0x67452301; + ctx->state[1] = 0xEFCDAB89; + ctx->state[2] = 0x98BADCFE; + ctx->state[3] = 0x10325476; + ctx->state[4] = 0xC3D2E1F0; + ctx->state[5] = 0x76543210; + ctx->state[6] = 0xFEDCBA98; + ctx->state[7] = 0x89ABCDEF; + ctx->state[8] = 0x01234567; + ctx->state[9] = 0x3C2D1E0F; + ctx->transform = ripemd320_transform; + break; + default: + return AVERROR(EINVAL); + } + ctx->count = 0; + return 0; +} + +void av_ripemd_update(AVRIPEMD* ctx, const uint8_t* data, unsigned int len) +{ + unsigned int i, j; + + j = ctx->count & 63; + ctx->count += len; +#if CONFIG_SMALL + for (i = 0; i < len; i++) { + ctx->buffer[j++] = data[i]; + if (64 == j) { + ctx->transform(ctx->state, ctx->buffer); + j = 0; + } + } +#else + if ((j + len) > 63) { + memcpy(&ctx->buffer[j], data, (i = 64 - j)); + ctx->transform(ctx->state, ctx->buffer); + for (; i + 63 < len; i += 64) + ctx->transform(ctx->state, &data[i]); + j = 0; + } else + i = 0; + memcpy(&ctx->buffer[j], &data[i], len - i); +#endif +} + +void av_ripemd_final(AVRIPEMD* ctx, uint8_t *digest) +{ + int i; + uint64_t finalcount = av_le2ne64(ctx->count << 3); + + av_ripemd_update(ctx, "\200", 1); + while ((ctx->count & 63) != 56) + av_ripemd_update(ctx, "", 1); + av_ripemd_update(ctx, (uint8_t *)&finalcount, 8); /* Should cause a transform() */ + for (i = 0; i < ctx->digest_len; i++) + AV_WL32(digest + i*4, ctx->state[i]); +} diff --git a/libavutil/ripemd.h b/libavutil/ripemd.h new file mode 100644 index 0000000000..7b0c8bc89c --- /dev/null +++ b/libavutil/ripemd.h @@ -0,0 +1,75 @@ +/* + * Copyright (C) 2007 Michael Niedermayer <michaelni@gmx.at> + * Copyright (C) 2013 James Almer <jamrial@gmail.com> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#ifndef AVUTIL_RIPEMD_H +#define AVUTIL_RIPEMD_H + +#include <stdint.h> + +#include "attributes.h" +#include "version.h" + +/** + * @defgroup lavu_ripemd RIPEMD + * @ingroup lavu_crypto + * @{ + */ + +extern const int av_ripemd_size; + +struct AVRIPEMD; + +/** + * Allocate an AVRIPEMD context. + */ +struct AVRIPEMD *av_ripemd_alloc(void); + +/** + * Initialize RIPEMD hashing. + * + * @param context pointer to the function context (of size av_ripemd_size) + * @param bits number of bits in digest (128, 160, 256 or 320 bits) + * @return zero if initialization succeeded, -1 otherwise + */ +int av_ripemd_init(struct AVRIPEMD* context, int bits); + +/** + * Update hash value. + * + * @param context hash function context + * @param data input data to update hash with + * @param len input data length + */ +void av_ripemd_update(struct AVRIPEMD* context, const uint8_t* data, unsigned int len); + +/** + * Finish hashing and output digest value. + * + * @param context hash function context + * @param digest buffer where output digest value is stored + */ +void av_ripemd_final(struct AVRIPEMD* context, uint8_t *digest); + +/** + * @} + */ + +#endif /* AVUTIL_RIPEMD_H */ diff --git a/libavutil/samplefmt.c b/libavutil/samplefmt.c index bff600447c..c605b5ebdf 100644 --- a/libavutil/samplefmt.c +++ b/libavutil/samplefmt.c @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -24,7 +24,7 @@ #include <string.h> typedef struct SampleFmtInfo { - const char *name; + char name[8]; int bits; int planar; enum AVSampleFormat altform; ///< planar<->packed alternative form @@ -61,6 +61,15 @@ enum AVSampleFormat av_get_sample_fmt(const char *name) return AV_SAMPLE_FMT_NONE; } +enum AVSampleFormat av_get_alt_sample_fmt(enum AVSampleFormat sample_fmt, int planar) +{ + if (sample_fmt < 0 || sample_fmt >= AV_SAMPLE_FMT_NB) + return AV_SAMPLE_FMT_NONE; + if (sample_fmt_info[sample_fmt].planar == planar) + return sample_fmt; + return sample_fmt_info[sample_fmt].altform; +} + enum AVSampleFormat av_get_packed_sample_fmt(enum AVSampleFormat sample_fmt) { if (sample_fmt < 0 || sample_fmt >= AV_SAMPLE_FMT_NB) @@ -149,14 +158,14 @@ int av_samples_fill_arrays(uint8_t **audio_data, int *linesize, if (buf_size < 0) return buf_size; - audio_data[0] = buf; + audio_data[0] = (uint8_t *)buf; for (ch = 1; planar && ch < nb_channels; ch++) audio_data[ch] = audio_data[ch-1] + line_size; if (linesize) *linesize = line_size; - return 0; + return buf_size; } int av_samples_alloc(uint8_t **audio_data, int *linesize, int nb_channels, @@ -181,7 +190,22 @@ int av_samples_alloc(uint8_t **audio_data, int *linesize, int nb_channels, av_samples_set_silence(audio_data, 0, nb_samples, nb_channels, sample_fmt); - return 0; + return size; +} + +int av_samples_alloc_array_and_samples(uint8_t ***audio_data, int *linesize, int nb_channels, + int nb_samples, enum AVSampleFormat sample_fmt, int align) +{ + int ret, nb_planes = av_sample_fmt_is_planar(sample_fmt) ? nb_channels : 1; + + *audio_data = av_calloc(nb_planes, sizeof(**audio_data)); + if (!*audio_data) + return AVERROR(ENOMEM); + ret = av_samples_alloc(*audio_data, linesize, nb_channels, + nb_samples, sample_fmt, align); + if (ret < 0) + av_freep(audio_data); + return ret; } int av_samples_copy(uint8_t **dst, uint8_t * const *src, int dst_offset, @@ -197,8 +221,13 @@ int av_samples_copy(uint8_t **dst, uint8_t * const *src, int dst_offset, dst_offset *= block_align; src_offset *= block_align; - for (i = 0; i < planes; i++) - memcpy(dst[i] + dst_offset, src[i] + src_offset, data_size); + if((dst[0] < src[0] ? src[0] - dst[0] : dst[0] - src[0]) >= data_size) { + for (i = 0; i < planes; i++) + memcpy(dst[i] + dst_offset, src[i] + src_offset, data_size); + } else { + for (i = 0; i < planes; i++) + memmove(dst[i] + dst_offset, src[i] + src_offset, data_size); + } return 0; } diff --git a/libavutil/samplefmt.h b/libavutil/samplefmt.h index a770746ce3..57da2784d2 100644 --- a/libavutil/samplefmt.h +++ b/libavutil/samplefmt.h @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -35,20 +35,17 @@ */ /** - * Audio Sample Formats + * Audio sample formats * - * @par - * The data described by the sample format is always in native-endian order. - * Sample values can be expressed by native C types, hence the lack of a signed - * 24-bit sample format even though it is a common raw audio data format. + * - The data described by the sample format is always in native-endian order. + * Sample values can be expressed by native C types, hence the lack of a signed + * 24-bit sample format even though it is a common raw audio data format. * - * @par - * The floating-point formats are based on full volume being in the range - * [-1.0, 1.0]. Any values outside this range are beyond full volume level. + * - The floating-point formats are based on full volume being in the range + * [-1.0, 1.0]. Any values outside this range are beyond full volume level. * - * @par - * The data layout as used in av_samples_fill_arrays() and elsewhere in Libav - * (such as AVFrame in libavcodec) is as follows: + * - The data layout as used in av_samples_fill_arrays() and elsewhere in FFmpeg + * (such as AVFrame in libavcodec) is as follows: * * @par * For planar sample formats, each audio channel is in a separate data plane, @@ -56,6 +53,7 @@ * planes must be the same size. For packed sample formats, only the first data * plane is used, and samples for each channel are interleaved. In this case, * linesize is the buffer size, in bytes, for the 1 plane. + * */ enum AVSampleFormat { AV_SAMPLE_FMT_NONE = -1, @@ -87,6 +85,14 @@ const char *av_get_sample_fmt_name(enum AVSampleFormat sample_fmt); enum AVSampleFormat av_get_sample_fmt(const char *name); /** + * Return the planar<->packed alternative form of the given sample format, or + * AV_SAMPLE_FMT_NONE on error. If the passed sample_fmt is already in the + * requested planar/packed format, the format returned is the same as the + * input. + */ +enum AVSampleFormat av_get_alt_sample_fmt(enum AVSampleFormat sample_fmt, int planar); + +/** * Get the packed alternative form of the given sample format. * * If the passed sample_fmt is already in packed format, the format returned is @@ -162,16 +168,20 @@ int av_samples_get_buffer_size(int *linesize, int nb_channels, int nb_samples, */ /** - * Fill channel data pointers and linesize for samples with sample + * Fill plane data pointers and linesize for samples with sample * format sample_fmt. * - * The pointers array is filled with the pointers to the samples data: + * The audio_data array is filled with the pointers to the samples data planes: * for planar, set the start point of each channel's data within the buffer, * for packed, set the start point of the entire buffer only. * - * The linesize array is filled with the aligned size of each channel's data - * buffer for planar layout, or the aligned size of the buffer for all channels - * for packed layout. + * The value pointed to by linesize is set to the aligned size of each + * channel's data buffer for planar layout, or to the aligned size of the + * buffer for all channels for packed layout. + * + * The buffer in buf must be big enough to contain all the samples + * (use av_samples_get_buffer_size() to compute its minimum size), + * otherwise the audio_data pointers will point to invalid data. * * @see enum AVSampleFormat * The documentation for AVSampleFormat describes the data layout. @@ -183,7 +193,9 @@ int av_samples_get_buffer_size(int *linesize, int nb_channels, int nb_samples, * @param nb_samples the number of samples in a single channel * @param sample_fmt the sample format * @param align buffer size alignment (0 = default, 1 = no alignment) - * @return 0 on success or a negative error code on failure + * @return >=0 on success or a negative error code on failure + * @todo return minimum size in bytes required for the buffer in case + * of success at the next bump */ int av_samples_fill_arrays(uint8_t **audio_data, int *linesize, const uint8_t *buf, @@ -204,13 +216,27 @@ int av_samples_fill_arrays(uint8_t **audio_data, int *linesize, * @param nb_channels number of audio channels * @param nb_samples number of samples per channel * @param align buffer size alignment (0 = default, 1 = no alignment) - * @return 0 on success or a negative error code on failure + * @return >=0 on success or a negative error code on failure + * @todo return the size of the allocated buffer in case of success at the next bump * @see av_samples_fill_arrays() + * @see av_samples_alloc_array_and_samples() */ int av_samples_alloc(uint8_t **audio_data, int *linesize, int nb_channels, int nb_samples, enum AVSampleFormat sample_fmt, int align); /** + * Allocate a data pointers array, samples buffer for nb_samples + * samples, and fill data pointers and linesize accordingly. + * + * This is the same as av_samples_alloc(), but also allocates the data + * pointers array. + * + * @see av_samples_alloc() + */ +int av_samples_alloc_array_and_samples(uint8_t ***audio_data, int *linesize, int nb_channels, + int nb_samples, enum AVSampleFormat sample_fmt, int align); + +/** * Copy samples from src to dst. * * @param dst destination array of pointers to data planes diff --git a/libavutil/sh4/bswap.h b/libavutil/sh4/bswap.h index 1ff1bfdd37..48dd27f806 100644 --- a/libavutil/sh4/bswap.h +++ b/libavutil/sh4/bswap.h @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/sha-test.c b/libavutil/sha-test.c index e43a5a53d5..1557591c75 100644 --- a/libavutil/sha-test.c +++ b/libavutil/sha-test.c @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -25,7 +25,7 @@ int main(void) int i, j, k; AVSHA ctx; unsigned char digest[32]; - const int lengths[3] = { 160, 224, 256 }; + static const int lengths[3] = { 160, 224, 256 }; for (j = 0; j < 3; j++) { printf("Testing SHA-%d\n", lengths[j]); diff --git a/libavutil/sha.c b/libavutil/sha.c index 404effade2..5b42ccf1b2 100644 --- a/libavutil/sha.c +++ b/libavutil/sha.c @@ -4,20 +4,20 @@ * based on public domain SHA-1 code by Steve Reid <steve@edmweb.com> * and on BSD-licensed SHA-2 code by Aaron D. Gifford * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -40,6 +40,8 @@ typedef struct AVSHA { void (*transform)(uint32_t *state, const uint8_t buffer[64]); } AVSHA; +const int av_sha_size = sizeof(AVSHA); + struct AVSHA *av_sha_alloc(void) { return av_mallocz(sizeof(struct AVSHA)); @@ -49,13 +51,13 @@ struct AVSHA *av_sha_alloc(void) /* (R0+R1), R2, R3, R4 are the different operations used in SHA1 */ #define blk0(i) (block[i] = AV_RB32(buffer + 4 * (i))) -#define blk(i) (block[i] = rol(block[i-3] ^ block[i-8] ^ block[i-14] ^ block[i-16], 1)) +#define blk(i) (block[i] = rol(block[(i)-3] ^ block[(i)-8] ^ block[(i)-14] ^ block[(i)-16], 1)) -#define R0(v,w,x,y,z,i) z += ((w&(x^y))^y) + blk0(i) + 0x5A827999 + rol(v, 5); w = rol(w, 30); -#define R1(v,w,x,y,z,i) z += ((w&(x^y))^y) + blk (i) + 0x5A827999 + rol(v, 5); w = rol(w, 30); -#define R2(v,w,x,y,z,i) z += ( w^x ^y) + blk (i) + 0x6ED9EBA1 + rol(v, 5); w = rol(w, 30); -#define R3(v,w,x,y,z,i) z += (((w|x)&y)|(w&x)) + blk (i) + 0x8F1BBCDC + rol(v, 5); w = rol(w, 30); -#define R4(v,w,x,y,z,i) z += ( w^x ^y) + blk (i) + 0xCA62C1D6 + rol(v, 5); w = rol(w, 30); +#define R0(v,w,x,y,z,i) z += (((w)&((x)^(y)))^(y)) + blk0(i) + 0x5A827999 + rol(v, 5); w = rol(w, 30); +#define R1(v,w,x,y,z,i) z += (((w)&((x)^(y)))^(y)) + blk (i) + 0x5A827999 + rol(v, 5); w = rol(w, 30); +#define R2(v,w,x,y,z,i) z += ( (w)^(x) ^(y)) + blk (i) + 0x6ED9EBA1 + rol(v, 5); w = rol(w, 30); +#define R3(v,w,x,y,z,i) z += ((((w)|(x))&(y))|((w)&(x))) + blk (i) + 0x8F1BBCDC + rol(v, 5); w = rol(w, 30); +#define R4(v,w,x,y,z,i) z += ( (w)^(x) ^(y)) + blk (i) + 0xCA62C1D6 + rol(v, 5); w = rol(w, 30); /* Hash a single 512-bit block. This is the core of the algorithm. */ @@ -96,39 +98,53 @@ static void sha1_transform(uint32_t state[5], const uint8_t buffer[64]) a = t; } #else - for (i = 0; i < 15; i += 5) { - R0(a, b, c, d, e, 0 + i); - R0(e, a, b, c, d, 1 + i); - R0(d, e, a, b, c, 2 + i); - R0(c, d, e, a, b, 3 + i); - R0(b, c, d, e, a, 4 + i); - } + +#define R1_0 \ + R0(a, b, c, d, e, 0 + i); \ + R0(e, a, b, c, d, 1 + i); \ + R0(d, e, a, b, c, 2 + i); \ + R0(c, d, e, a, b, 3 + i); \ + R0(b, c, d, e, a, 4 + i); \ + i += 5 + + i = 0; + R1_0; R1_0; R1_0; R0(a, b, c, d, e, 15); R1(e, a, b, c, d, 16); R1(d, e, a, b, c, 17); R1(c, d, e, a, b, 18); R1(b, c, d, e, a, 19); - for (i = 20; i < 40; i += 5) { - R2(a, b, c, d, e, 0 + i); - R2(e, a, b, c, d, 1 + i); - R2(d, e, a, b, c, 2 + i); - R2(c, d, e, a, b, 3 + i); - R2(b, c, d, e, a, 4 + i); - } - for (; i < 60; i += 5) { - R3(a, b, c, d, e, 0 + i); - R3(e, a, b, c, d, 1 + i); - R3(d, e, a, b, c, 2 + i); - R3(c, d, e, a, b, 3 + i); - R3(b, c, d, e, a, 4 + i); - } - for (; i < 80; i += 5) { - R4(a, b, c, d, e, 0 + i); - R4(e, a, b, c, d, 1 + i); - R4(d, e, a, b, c, 2 + i); - R4(c, d, e, a, b, 3 + i); - R4(b, c, d, e, a, 4 + i); - } + +#define R1_20 \ + R2(a, b, c, d, e, 0 + i); \ + R2(e, a, b, c, d, 1 + i); \ + R2(d, e, a, b, c, 2 + i); \ + R2(c, d, e, a, b, 3 + i); \ + R2(b, c, d, e, a, 4 + i); \ + i += 5 + + i = 20; + R1_20; R1_20; R1_20; R1_20; + +#define R1_40 \ + R3(a, b, c, d, e, 0 + i); \ + R3(e, a, b, c, d, 1 + i); \ + R3(d, e, a, b, c, 2 + i); \ + R3(c, d, e, a, b, 3 + i); \ + R3(b, c, d, e, a, 4 + i); \ + i += 5 + + R1_40; R1_40; R1_40; R1_40; + +#define R1_60 \ + R4(a, b, c, d, e, 0 + i); \ + R4(e, a, b, c, d, 1 + i); \ + R4(d, e, a, b, c, 2 + i); \ + R4(c, d, e, a, b, 3 + i); \ + R4(b, c, d, e, a, 4 + i); \ + i += 5 + + R1_60; R1_60; R1_60; R1_60; #endif state[0] += a; state[1] += b; @@ -158,7 +174,7 @@ static const uint32_t K256[64] = { #define Ch(x,y,z) (((x) & ((y) ^ (z))) ^ (z)) -#define Maj(x,y,z) ((((x) | (y)) & (z)) | ((x) & (y))) +#define Maj(z,y,x) ((((x) | (y)) & (z)) | ((x) & (y))) #define Sigma0_256(x) (rol((x), 30) ^ rol((x), 19) ^ rol((x), 10)) #define Sigma1_256(x) (rol((x), 26) ^ rol((x), 21) ^ rol((x), 7)) @@ -216,27 +232,32 @@ static void sha256_transform(uint32_t *state, const uint8_t buffer[64]) a = T1 + T2; } #else - for (i = 0; i < 16;) { - ROUND256_0_TO_15(a, b, c, d, e, f, g, h); - ROUND256_0_TO_15(h, a, b, c, d, e, f, g); - ROUND256_0_TO_15(g, h, a, b, c, d, e, f); - ROUND256_0_TO_15(f, g, h, a, b, c, d, e); - ROUND256_0_TO_15(e, f, g, h, a, b, c, d); - ROUND256_0_TO_15(d, e, f, g, h, a, b, c); - ROUND256_0_TO_15(c, d, e, f, g, h, a, b); - ROUND256_0_TO_15(b, c, d, e, f, g, h, a); - } - for (; i < 64;) { - ROUND256_16_TO_63(a, b, c, d, e, f, g, h); - ROUND256_16_TO_63(h, a, b, c, d, e, f, g); - ROUND256_16_TO_63(g, h, a, b, c, d, e, f); - ROUND256_16_TO_63(f, g, h, a, b, c, d, e); - ROUND256_16_TO_63(e, f, g, h, a, b, c, d); - ROUND256_16_TO_63(d, e, f, g, h, a, b, c); - ROUND256_16_TO_63(c, d, e, f, g, h, a, b); - ROUND256_16_TO_63(b, c, d, e, f, g, h, a); - } + i = 0; +#define R256_0 \ + ROUND256_0_TO_15(a, b, c, d, e, f, g, h); \ + ROUND256_0_TO_15(h, a, b, c, d, e, f, g); \ + ROUND256_0_TO_15(g, h, a, b, c, d, e, f); \ + ROUND256_0_TO_15(f, g, h, a, b, c, d, e); \ + ROUND256_0_TO_15(e, f, g, h, a, b, c, d); \ + ROUND256_0_TO_15(d, e, f, g, h, a, b, c); \ + ROUND256_0_TO_15(c, d, e, f, g, h, a, b); \ + ROUND256_0_TO_15(b, c, d, e, f, g, h, a) + + R256_0; R256_0; + +#define R256_16 \ + ROUND256_16_TO_63(a, b, c, d, e, f, g, h); \ + ROUND256_16_TO_63(h, a, b, c, d, e, f, g); \ + ROUND256_16_TO_63(g, h, a, b, c, d, e, f); \ + ROUND256_16_TO_63(f, g, h, a, b, c, d, e); \ + ROUND256_16_TO_63(e, f, g, h, a, b, c, d); \ + ROUND256_16_TO_63(d, e, f, g, h, a, b, c); \ + ROUND256_16_TO_63(c, d, e, f, g, h, a, b); \ + ROUND256_16_TO_63(b, c, d, e, f, g, h, a) + + R256_16; R256_16; R256_16; + R256_16; R256_16; R256_16; #endif state[0] += a; state[1] += b; @@ -284,7 +305,7 @@ av_cold int av_sha_init(AVSHA *ctx, int bits) ctx->transform = sha256_transform; break; default: - return -1; + return AVERROR(EINVAL); } ctx->count = 0; return 0; diff --git a/libavutil/sha.h b/libavutil/sha.h index 86ea0b065e..bf4377e51b 100644 --- a/libavutil/sha.h +++ b/libavutil/sha.h @@ -1,20 +1,20 @@ /* * Copyright (C) 2007 Michael Niedermayer <michaelni@gmx.at> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -32,6 +32,8 @@ * @{ */ +extern const int av_sha_size; + struct AVSHA; /** diff --git a/libavutil/sha512-test.c b/libavutil/sha512-test.c new file mode 100644 index 0000000000..dd472aa868 --- /dev/null +++ b/libavutil/sha512-test.c @@ -0,0 +1,84 @@ +/* + * Copyright (C) 2007 Michael Niedermayer <michaelni@gmx.at> + * Copyright (C) 2009 Konstantin Shishkov + * Copyright (C) 2013 James Almer + * based on BSD-licensed SHA-2 code by Aaron D. Gifford + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "sha512.c" + +#include <stdio.h> + +int main(void) +{ + int i, j, k; + AVSHA512 ctx; + unsigned char digest[64]; + static const int lengths[4] = { 224, 256, 384, 512 }; + + for (j = 0; j < 4; j++) { + if (j < 2) printf("Testing SHA-512/%d\n", lengths[j]); + else printf("Testing SHA-%d\n", lengths[j]); + for (k = 0; k < 3; k++) { + av_sha512_init(&ctx, lengths[j]); + if (k == 0) + av_sha512_update(&ctx, "abc", 3); + else if (k == 1) + av_sha512_update(&ctx, "abcdefghbcdefghicdefghijdefghijkefghijklfghijklmghijklmn" + "hijklmnoijklmnopjklmnopqklmnopqrlmnopqrsmnopqrstnopqrstu", 112); + else + for (i = 0; i < 1000*1000; i++) + av_sha512_update(&ctx, "a", 1); + av_sha512_final(&ctx, digest); + for (i = 0; i < lengths[j] >> 3; i++) + printf("%02X", digest[i]); + putchar('\n'); + } + switch (j) { //test vectors (from FIPS PUB 180-4 Apendix A) + case 0: + printf("4634270f 707b6a54 daae7530 460842e2 0e37ed26 5ceee9a4 3e8924aa\n" + "23fec5bb 94d60b23 30819264 0b0c4533 35d66473 4fe40e72 68674af9\n" + "37ab331d 76f0d36d e422bd0e deb22a28 accd487b 7a8453ae 965dd287\n"); + break; + case 1: + printf("53048e26 81941ef9 9b2e29b7 6b4c7dab e4c2d0c6 34fc6d46 e0e2f131 07e7af23\n" + "3928e184 fb8690f8 40da3988 121d31be 65cb9d3e f83ee614 6feac861 e19b563a\n" + "9a59a052 930187a9 7038cae6 92f30708 aa649192 3ef51943 94dc68d5 6c74fb21\n"); + break; + case 2: + printf("cb00753f 45a35e8b b5a03d69 9ac65007 272c32ab 0eded163 " + "1a8b605a 43ff5bed 8086072b a1e7cc23 58baeca1 34c825a7\n" + "09330c33 f71147e8 3d192fc7 82cd1b47 53111b17 3b3b05d2 " + "2fa08086 e3b0f712 fcc7c71a 557e2db9 66c3e9fa 91746039\n" + "9d0e1809 716474cb 086e834e 310a4a1c ed149e9c 00f24852 " + "7972cec5 704c2a5b 07b8b3dc 38ecc4eb ae97ddd8 7f3d8985\n"); + break; + case 3: + printf("ddaf35a1 93617aba cc417349 ae204131 12e6fa4e 89a97ea2 0a9eeee6 4b55d39a " + "2192992a 274fc1a8 36ba3c23 a3feebbd 454d4423 643ce80e 2a9ac94f a54ca49f\n" + "8e959b75 dae313da 8cf4f728 14fc143f 8f7779c6 eb9f7fa1 7299aead b6889018 " + "501d289e 4900f7e4 331b99de c4b5433a c7d329ee b6dd2654 5e96e55b 874be909\n" + "e718483d 0ce76964 4e2e42c7 bc15b463 8e1f98b1 3b204428 5632a803 afa973eb " + "de0ff244 877ea60a 4cb0432c e577c31b eb009c5c 2c49aa2e 4eadb217 ad8cc09b\n"); + break; + } + } + + return 0; +} diff --git a/libavutil/sha512.c b/libavutil/sha512.c new file mode 100644 index 0000000000..2c5da25591 --- /dev/null +++ b/libavutil/sha512.c @@ -0,0 +1,283 @@ +/* + * Copyright (C) 2007 Michael Niedermayer <michaelni@gmx.at> + * Copyright (C) 2009 Konstantin Shishkov + * Copyright (C) 2013 James Almer + * based on BSD-licensed SHA-2 code by Aaron D. Gifford + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include <string.h> + +#include "attributes.h" +#include "avutil.h" +#include "bswap.h" +#include "sha512.h" +#include "intreadwrite.h" +#include "mem.h" + +/** hash context */ +typedef struct AVSHA512 { + uint8_t digest_len; ///< digest length in 64-bit words + uint64_t count; ///< number of bytes in buffer + uint8_t buffer[128]; ///< 1024-bit buffer of input values used in hash updating + uint64_t state[8]; ///< current hash value +} AVSHA512; + +const int av_sha512_size = sizeof(AVSHA512); + +struct AVSHA512 *av_sha512_alloc(void) +{ + return av_mallocz(sizeof(struct AVSHA512)); +} + +static const uint64_t K512[80] = { + UINT64_C(0x428a2f98d728ae22), UINT64_C(0x7137449123ef65cd), + UINT64_C(0xb5c0fbcfec4d3b2f), UINT64_C(0xe9b5dba58189dbbc), + UINT64_C(0x3956c25bf348b538), UINT64_C(0x59f111f1b605d019), + UINT64_C(0x923f82a4af194f9b), UINT64_C(0xab1c5ed5da6d8118), + UINT64_C(0xd807aa98a3030242), UINT64_C(0x12835b0145706fbe), + UINT64_C(0x243185be4ee4b28c), UINT64_C(0x550c7dc3d5ffb4e2), + UINT64_C(0x72be5d74f27b896f), UINT64_C(0x80deb1fe3b1696b1), + UINT64_C(0x9bdc06a725c71235), UINT64_C(0xc19bf174cf692694), + UINT64_C(0xe49b69c19ef14ad2), UINT64_C(0xefbe4786384f25e3), + UINT64_C(0x0fc19dc68b8cd5b5), UINT64_C(0x240ca1cc77ac9c65), + UINT64_C(0x2de92c6f592b0275), UINT64_C(0x4a7484aa6ea6e483), + UINT64_C(0x5cb0a9dcbd41fbd4), UINT64_C(0x76f988da831153b5), + UINT64_C(0x983e5152ee66dfab), UINT64_C(0xa831c66d2db43210), + UINT64_C(0xb00327c898fb213f), UINT64_C(0xbf597fc7beef0ee4), + UINT64_C(0xc6e00bf33da88fc2), UINT64_C(0xd5a79147930aa725), + UINT64_C(0x06ca6351e003826f), UINT64_C(0x142929670a0e6e70), + UINT64_C(0x27b70a8546d22ffc), UINT64_C(0x2e1b21385c26c926), + UINT64_C(0x4d2c6dfc5ac42aed), UINT64_C(0x53380d139d95b3df), + UINT64_C(0x650a73548baf63de), UINT64_C(0x766a0abb3c77b2a8), + UINT64_C(0x81c2c92e47edaee6), UINT64_C(0x92722c851482353b), + UINT64_C(0xa2bfe8a14cf10364), UINT64_C(0xa81a664bbc423001), + UINT64_C(0xc24b8b70d0f89791), UINT64_C(0xc76c51a30654be30), + UINT64_C(0xd192e819d6ef5218), UINT64_C(0xd69906245565a910), + UINT64_C(0xf40e35855771202a), UINT64_C(0x106aa07032bbd1b8), + UINT64_C(0x19a4c116b8d2d0c8), UINT64_C(0x1e376c085141ab53), + UINT64_C(0x2748774cdf8eeb99), UINT64_C(0x34b0bcb5e19b48a8), + UINT64_C(0x391c0cb3c5c95a63), UINT64_C(0x4ed8aa4ae3418acb), + UINT64_C(0x5b9cca4f7763e373), UINT64_C(0x682e6ff3d6b2b8a3), + UINT64_C(0x748f82ee5defb2fc), UINT64_C(0x78a5636f43172f60), + UINT64_C(0x84c87814a1f0ab72), UINT64_C(0x8cc702081a6439ec), + UINT64_C(0x90befffa23631e28), UINT64_C(0xa4506cebde82bde9), + UINT64_C(0xbef9a3f7b2c67915), UINT64_C(0xc67178f2e372532b), + UINT64_C(0xca273eceea26619c), UINT64_C(0xd186b8c721c0c207), + UINT64_C(0xeada7dd6cde0eb1e), UINT64_C(0xf57d4f7fee6ed178), + UINT64_C(0x06f067aa72176fba), UINT64_C(0x0a637dc5a2c898a6), + UINT64_C(0x113f9804bef90dae), UINT64_C(0x1b710b35131c471b), + UINT64_C(0x28db77f523047d84), UINT64_C(0x32caab7b40c72493), + UINT64_C(0x3c9ebe0a15c9bebc), UINT64_C(0x431d67c49c100d4c), + UINT64_C(0x4cc5d4becb3e42b6), UINT64_C(0x597f299cfc657e2a), + UINT64_C(0x5fcb6fab3ad6faec), UINT64_C(0x6c44198c4a475817), +}; + +#define ror(value, bits) (((value) >> (bits)) | ((value) << (64 - (bits)))) + +#define Ch(x,y,z) (((x) & ((y) ^ (z))) ^ (z)) +#define Maj(z,y,x) ((((x) | (y)) & (z)) | ((x) & (y))) + +#define Sigma0_512(x) (ror((x), 28) ^ ror((x), 34) ^ ror((x), 39)) +#define Sigma1_512(x) (ror((x), 14) ^ ror((x), 18) ^ ror((x), 41)) +#define sigma0_512(x) (ror((x), 1) ^ ror((x), 8) ^ ((x) >> 7)) +#define sigma1_512(x) (ror((x), 19) ^ ror((x), 61) ^ ((x) >> 6)) + +#define blk0(i) (block[i] = AV_RB64(buffer + 8 * (i))) +#define blk(i) (block[i] = block[i - 16] + sigma0_512(block[i - 15]) + \ + sigma1_512(block[i - 2]) + block[i - 7]) + +#define ROUND512(a,b,c,d,e,f,g,h) \ + T1 += (h) + Sigma1_512(e) + Ch((e), (f), (g)) + K512[i]; \ + (d) += T1; \ + (h) = T1 + Sigma0_512(a) + Maj((a), (b), (c)); \ + i++ + +#define ROUND512_0_TO_15(a,b,c,d,e,f,g,h) \ + T1 = blk0(i); \ + ROUND512(a,b,c,d,e,f,g,h) + +#define ROUND512_16_TO_80(a,b,c,d,e,f,g,h) \ + T1 = blk(i); \ + ROUND512(a,b,c,d,e,f,g,h) + +static void sha512_transform(uint64_t *state, const uint8_t buffer[128]) +{ + uint64_t a, b, c, d, e, f, g, h; + uint64_t block[80]; + uint64_t T1; + int i; + + a = state[0]; + b = state[1]; + c = state[2]; + d = state[3]; + e = state[4]; + f = state[5]; + g = state[6]; + h = state[7]; +#if CONFIG_SMALL + for (i = 0; i < 80; i++) { + uint64_t T2; + if (i < 16) + T1 = blk0(i); + else + T1 = blk(i); + T1 += h + Sigma1_512(e) + Ch(e, f, g) + K512[i]; + T2 = Sigma0_512(a) + Maj(a, b, c); + h = g; + g = f; + f = e; + e = d + T1; + d = c; + c = b; + b = a; + a = T1 + T2; + } +#else + +#define R512_0 \ + ROUND512_0_TO_15(a, b, c, d, e, f, g, h); \ + ROUND512_0_TO_15(h, a, b, c, d, e, f, g); \ + ROUND512_0_TO_15(g, h, a, b, c, d, e, f); \ + ROUND512_0_TO_15(f, g, h, a, b, c, d, e); \ + ROUND512_0_TO_15(e, f, g, h, a, b, c, d); \ + ROUND512_0_TO_15(d, e, f, g, h, a, b, c); \ + ROUND512_0_TO_15(c, d, e, f, g, h, a, b); \ + ROUND512_0_TO_15(b, c, d, e, f, g, h, a) + + i = 0; + R512_0; R512_0; + +#define R512_16 \ + ROUND512_16_TO_80(a, b, c, d, e, f, g, h); \ + ROUND512_16_TO_80(h, a, b, c, d, e, f, g); \ + ROUND512_16_TO_80(g, h, a, b, c, d, e, f); \ + ROUND512_16_TO_80(f, g, h, a, b, c, d, e); \ + ROUND512_16_TO_80(e, f, g, h, a, b, c, d); \ + ROUND512_16_TO_80(d, e, f, g, h, a, b, c); \ + ROUND512_16_TO_80(c, d, e, f, g, h, a, b); \ + ROUND512_16_TO_80(b, c, d, e, f, g, h, a) + + R512_16; R512_16; R512_16; R512_16; + R512_16; R512_16; R512_16; R512_16; +#endif + state[0] += a; + state[1] += b; + state[2] += c; + state[3] += d; + state[4] += e; + state[5] += f; + state[6] += g; + state[7] += h; +} + + +av_cold int av_sha512_init(AVSHA512 *ctx, int bits) +{ + ctx->digest_len = bits >> 6; + switch (bits) { + case 224: // SHA-512/224 + ctx->state[0] = UINT64_C(0x8C3D37C819544DA2); + ctx->state[1] = UINT64_C(0x73E1996689DCD4D6); + ctx->state[2] = UINT64_C(0x1DFAB7AE32FF9C82); + ctx->state[3] = UINT64_C(0x679DD514582F9FCF); + ctx->state[4] = UINT64_C(0x0F6D2B697BD44DA8); + ctx->state[5] = UINT64_C(0x77E36F7304C48942); + ctx->state[6] = UINT64_C(0x3F9D85A86A1D36C8); + ctx->state[7] = UINT64_C(0x1112E6AD91D692A1); + break; + case 256: // SHA-512/256 + ctx->state[0] = UINT64_C(0x22312194FC2BF72C); + ctx->state[1] = UINT64_C(0x9F555FA3C84C64C2); + ctx->state[2] = UINT64_C(0x2393B86B6F53B151); + ctx->state[3] = UINT64_C(0x963877195940EABD); + ctx->state[4] = UINT64_C(0x96283EE2A88EFFE3); + ctx->state[5] = UINT64_C(0xBE5E1E2553863992); + ctx->state[6] = UINT64_C(0x2B0199FC2C85B8AA); + ctx->state[7] = UINT64_C(0x0EB72DDC81C52CA2); + break; + case 384: // SHA-384 + ctx->state[0] = UINT64_C(0xCBBB9D5DC1059ED8); + ctx->state[1] = UINT64_C(0x629A292A367CD507); + ctx->state[2] = UINT64_C(0x9159015A3070DD17); + ctx->state[3] = UINT64_C(0x152FECD8F70E5939); + ctx->state[4] = UINT64_C(0x67332667FFC00B31); + ctx->state[5] = UINT64_C(0x8EB44A8768581511); + ctx->state[6] = UINT64_C(0xDB0C2E0D64F98FA7); + ctx->state[7] = UINT64_C(0x47B5481DBEFA4FA4); + break; + case 512: // SHA-512 + ctx->state[0] = UINT64_C(0x6A09E667F3BCC908); + ctx->state[1] = UINT64_C(0xBB67AE8584CAA73B); + ctx->state[2] = UINT64_C(0x3C6EF372FE94F82B); + ctx->state[3] = UINT64_C(0xA54FF53A5F1D36F1); + ctx->state[4] = UINT64_C(0x510E527FADE682D1); + ctx->state[5] = UINT64_C(0x9B05688C2B3E6C1F); + ctx->state[6] = UINT64_C(0x1F83D9ABFB41BD6B); + ctx->state[7] = UINT64_C(0x5BE0CD19137E2179); + break; + default: + return AVERROR(EINVAL); + } + ctx->count = 0; + return 0; +} + +void av_sha512_update(AVSHA512* ctx, const uint8_t* data, unsigned int len) +{ + unsigned int i, j; + + j = ctx->count & 127; + ctx->count += len; +#if CONFIG_SMALL + for (i = 0; i < len; i++) { + ctx->buffer[j++] = data[i]; + if (128 == j) { + sha512_transform(ctx->state, ctx->buffer); + j = 0; + } + } +#else + if ((j + len) > 127) { + memcpy(&ctx->buffer[j], data, (i = 128 - j)); + sha512_transform(ctx->state, ctx->buffer); + for (; i + 127 < len; i += 128) + sha512_transform(ctx->state, &data[i]); + j = 0; + } else + i = 0; + memcpy(&ctx->buffer[j], &data[i], len - i); +#endif +} + +void av_sha512_final(AVSHA512* ctx, uint8_t *digest) +{ + uint64_t i = 0; + uint64_t finalcount = av_be2ne64(ctx->count << 3); + + av_sha512_update(ctx, "\200", 1); + while ((ctx->count & 127) != 112) + av_sha512_update(ctx, "", 1); + av_sha512_update(ctx, (uint8_t *)&i, 8); + av_sha512_update(ctx, (uint8_t *)&finalcount, 8); /* Should cause a transform() */ + for (i = 0; i < ctx->digest_len; i++) + AV_WB64(digest + i*8, ctx->state[i]); + if (ctx->digest_len & 1) /* SHA512/224 is 28 bytes, and is not divisible by 8. */ + AV_WB32(digest + i*8, ctx->state[i] >> 32); +} diff --git a/libavutil/sha512.h b/libavutil/sha512.h new file mode 100644 index 0000000000..7b08701477 --- /dev/null +++ b/libavutil/sha512.h @@ -0,0 +1,75 @@ +/* + * Copyright (C) 2007 Michael Niedermayer <michaelni@gmx.at> + * Copyright (C) 2013 James Almer <jamrial@gmail.com> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#ifndef AVUTIL_SHA512_H +#define AVUTIL_SHA512_H + +#include <stdint.h> + +#include "attributes.h" +#include "version.h" + +/** + * @defgroup lavu_sha512 SHA512 + * @ingroup lavu_crypto + * @{ + */ + +extern const int av_sha512_size; + +struct AVSHA512; + +/** + * Allocate an AVSHA512 context. + */ +struct AVSHA512 *av_sha512_alloc(void); + +/** + * Initialize SHA-2 512 hashing. + * + * @param context pointer to the function context (of size av_sha512_size) + * @param bits number of bits in digest (224, 256, 384 or 512 bits) + * @return zero if initialization succeeded, -1 otherwise + */ +int av_sha512_init(struct AVSHA512* context, int bits); + +/** + * Update hash value. + * + * @param context hash function context + * @param data input data to update hash with + * @param len input data length + */ +void av_sha512_update(struct AVSHA512* context, const uint8_t* data, unsigned int len); + +/** + * Finish hashing and output digest value. + * + * @param context hash function context + * @param digest buffer where output digest value is stored + */ +void av_sha512_final(struct AVSHA512* context, uint8_t *digest); + +/** + * @} + */ + +#endif /* AVUTIL_SHA512_H */ diff --git a/libavutil/softfloat-test.c b/libavutil/softfloat-test.c new file mode 100644 index 0000000000..3293450df8 --- /dev/null +++ b/libavutil/softfloat-test.c @@ -0,0 +1,155 @@ +/* + * copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include <inttypes.h> +#include "softfloat.h" +#include "common.h" +#include "log.h" + +#include <stdio.h> + +static const SoftFloat FLOAT_0_017776489257 = {0x1234, 12}; +static const SoftFloat FLOAT_1374_40625 = {0xabcd, 25}; +static const SoftFloat FLOAT_0_1249694824218 = {0xFFF, 15}; + + +int main(void){ + SoftFloat one= av_int2sf(1, 0); + SoftFloat sf1, sf2, sf3; + double d1, d2, d3; + int i, j; + av_log_set_level(AV_LOG_DEBUG); + + d1= 1; + for(i= 0; i<10; i++){ + d1= 1/(d1+1); + } + printf("test1 double=%d\n", (int)(d1 * (1<<24))); + + sf1= one; + for(i= 0; i<10; i++){ + sf1= av_div_sf(one, av_normalize_sf(av_add_sf(one, sf1))); + } + printf("test1 sf =%d\n", av_sf2int(sf1, 24)); + + + for(i= 0; i<100; i++){ + START_TIMER + d1= i; + d2= i/100.0; + for(j= 0; j<1000; j++){ + d1= (d1+1)*d2; + } + STOP_TIMER("float add mul") + } + printf("test2 double=%d\n", (int)(d1 * (1<<24))); + + for(i= 0; i<100; i++){ + START_TIMER + sf1= av_int2sf(i, 0); + sf2= av_div_sf(av_int2sf(i, 2), av_int2sf(200, 3)); + for(j= 0; j<1000; j++){ + sf1= av_mul_sf(av_add_sf(sf1, one),sf2); + } + STOP_TIMER("softfloat add mul") + } + printf("test2 sf =%d (%d %d)\n", av_sf2int(sf1, 24), sf1.exp, sf1.mant); + + d1 = 0.0177764893; + d2 = 1374.40625; + d3 = 0.1249694824; + d2 += d1; + d3 += d2; + printf("test3 double: %.10lf\n", d3); + + sf1 = FLOAT_0_017776489257; + sf2 = FLOAT_1374_40625; + sf3 = FLOAT_0_1249694824218; + sf2 = av_add_sf(sf1, sf2); + sf3 = av_add_sf(sf3, sf2); + printf("test3 softfloat: %.10lf (0x%08x %d)\n", (double)av_sf2double(sf3), sf3.mant, sf3.exp); + + sf1 = av_int2sf(0xFFFFFFF0, 0); + printf("test4 softfloat: %.10lf (0x%08x %d)\n", (double)av_sf2double(sf1), sf1.mant, sf1.exp); + sf1 = av_int2sf(0x00000010, 0); + printf("test4 softfloat: %.10lf (0x%08x %d)\n", (double)av_sf2double(sf1), sf1.mant, sf1.exp); + + sf1 = av_int2sf(0x1FFFFFFF, 0); + printf("test4 softfloat: %.10lf (0x%08x %d)\n", (double)av_sf2double(sf1), sf1.mant, sf1.exp); + sf1 = av_int2sf(0xE0000001, 0); + printf("test4 softfloat: %.10lf (0x%08x %d)\n", (double)av_sf2double(sf1), sf1.mant, sf1.exp); + + + sf1 = (SoftFloat){ 0x20000000, MIN_EXP }; + sf1 = av_mul_sf(sf1, sf1); + printf("test5 softfloat: %.10lf (0x%08x %d)\n", (double)av_sf2double(sf1), sf1.mant, sf1.exp); + + sf1 = (SoftFloat){ 0x20000000, MIN_EXP }; + sf2 = (SoftFloat){ 0x20000000, MAX_EXP }; + i = av_cmp_sf(sf1, sf2); + j = av_cmp_sf(sf2, sf1); + sf1 = av_div_sf(sf1, sf2); + printf("test6 softfloat: %.10lf (0x%08x %d) %d %d\n", (double)av_sf2double(sf1), sf1.mant, sf1.exp, i, j); + + for(i= -50; i<50; i++) { + sf1= av_int2sf(i, 0); + for(j= -50; j<50; j++) { + int c; + sf2= av_int2sf(j, 0); + c = av_cmp_sf(sf1, sf2); + if (FFDIFFSIGN(i,j) != c && (FFDIFFSIGN(i,j)^c)<0) { + printf("av_cmp_sf failed at %d %d as %X\n", i, j, c); + } + c = av_gt_sf(sf1, sf2); + if ((i>j) != c) { + printf("av_gt_sf failed at %d %d as %X\n", i, j, c); + } + } + sf1 = av_int2sf(1, i); + for(j = -50; j < 50; j++) { + int c; + sf2 = av_int2sf(1, j); + c = av_cmp_sf(sf2, sf1); + if (FFDIFFSIGN(i,j) != c && (FFDIFFSIGN(i,j)^c) < 0) { + printf("av_cmp_sf failed2 at %d %d as %X\n", i, j, c); + } + c = av_gt_sf(sf1, sf2); + if ((i<j) != c) { + printf("av_gt_sf failed2 at %d %d as %X\n", i, j, c); + } + } + } + + + for(i= 0; i<4*36; i++){ + int s, c; + double errs, errc; + + av_sincos_sf(i*(1ULL<<32)/36/4, &s, &c); + errs = (double)s/ (1<<30) - sin(i*M_PI/36); + errc = (double)c/ (1<<30) - cos(i*M_PI/36); + if (fabs(errs) > 0.00000002 || fabs(errc) >0.001) { + printf("sincos FAIL %d %f %f %f %f\n", i, (float)s/ (1<<30), (float)c/ (1<<30), sin(i*M_PI/36), cos(i*M_PI/36)); + } + + } + return 0; + +} diff --git a/libavutil/softfloat.h b/libavutil/softfloat.h new file mode 100644 index 0000000000..4b895f014b --- /dev/null +++ b/libavutil/softfloat.h @@ -0,0 +1,259 @@ +/* + * Copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#ifndef AVUTIL_SOFTFLOAT_H +#define AVUTIL_SOFTFLOAT_H + +#include <stdint.h> +#include "common.h" + +#include "avassert.h" +#include "softfloat_tables.h" + +#define MIN_EXP -126 +#define MAX_EXP 126 +#define ONE_BITS 29 + +typedef struct SoftFloat{ + int32_t mant; + int32_t exp; +}SoftFloat; + +static const SoftFloat FLOAT_0 = { 0, MIN_EXP}; +static const SoftFloat FLOAT_05 = { 0x20000000, 0}; +static const SoftFloat FLOAT_1 = { 0x20000000, 1}; +static const SoftFloat FLOAT_EPSILON = { 0x29F16B12, -16}; +static const SoftFloat FLOAT_1584893192 = { 0x32B771ED, 1}; +static const SoftFloat FLOAT_100000 = { 0x30D40000, 17}; +static const SoftFloat FLOAT_0999999 = { 0x3FFFFBCE, 0}; + +static inline av_const double av_sf2double(SoftFloat v) { + v.exp -= ONE_BITS +1; + if(v.exp > 0) return (double)v.mant * (double)(1 << v.exp); + else return (double)v.mant / (double)(1 << (-v.exp)); +} + +static av_const SoftFloat av_normalize_sf(SoftFloat a){ + if(a.mant){ +#if 1 + while((a.mant + 0x1FFFFFFFU)<0x3FFFFFFFU){ + a.mant += a.mant; + a.exp -= 1; + } +#else + int s=ONE_BITS - av_log2(FFABS(a.mant)); + a.exp -= s; + a.mant <<= s; +#endif + if(a.exp < MIN_EXP){ + a.exp = MIN_EXP; + a.mant= 0; + } + }else{ + a.exp= MIN_EXP; + } + return a; +} + +static inline av_const SoftFloat av_normalize1_sf(SoftFloat a){ +#if 1 + if((int32_t)(a.mant + 0x40000000U) <= 0){ + a.exp++; + a.mant>>=1; + } + av_assert2(a.mant < 0x40000000 && a.mant > -0x40000000); + av_assert2(a.exp <= MAX_EXP); + return a; +#elif 1 + int t= a.mant + 0x40000000 < 0; + return (SoftFloat){ a.mant>>t, a.exp+t}; +#else + int t= (a.mant + 0x3FFFFFFFU)>>31; + return (SoftFloat){a.mant>>t, a.exp+t}; +#endif +} + +/** + * @return Will not be more denormalized than a*b. So if either input is + * normalized, then the output will not be worse then the other input. + * If both are normalized, then the output will be normalized. + */ +static inline av_const SoftFloat av_mul_sf(SoftFloat a, SoftFloat b){ + a.exp += b.exp; + av_assert2((int32_t)((a.mant * (int64_t)b.mant) >> ONE_BITS) == (a.mant * (int64_t)b.mant) >> ONE_BITS); + a.mant = (a.mant * (int64_t)b.mant) >> ONE_BITS; + a = av_normalize1_sf((SoftFloat){a.mant, a.exp - 1}); + if (!a.mant || a.exp < MIN_EXP) + return FLOAT_0; + return a; +} + +/** + * b has to be normalized and not zero. + * @return Will not be more denormalized than a. + */ +static inline av_const SoftFloat av_div_sf(SoftFloat a, SoftFloat b){ + a.exp -= b.exp; + a.mant = ((int64_t)a.mant<<(ONE_BITS+1)) / b.mant; + a = av_normalize1_sf(a); + if (!a.mant || a.exp < MIN_EXP) + return FLOAT_0; + return a; +} + +static inline av_const int av_cmp_sf(SoftFloat a, SoftFloat b){ + int t= a.exp - b.exp; + if (t <-31) return - b.mant ; + else if (t < 0) return (a.mant >> (-t)) - b.mant ; + else if (t < 32) return a.mant - (b.mant >> t); + else return a.mant ; +} + +static inline av_const int av_gt_sf(SoftFloat a, SoftFloat b) +{ + int t= a.exp - b.exp; + if (t <-31) return 0 > b.mant ; + else if (t < 0) return (a.mant >> (-t)) > b.mant ; + else if (t < 32) return a.mant > (b.mant >> t); + else return a.mant > 0 ; +} + +static inline av_const SoftFloat av_add_sf(SoftFloat a, SoftFloat b){ + int t= a.exp - b.exp; + if (t <-31) return b; + else if (t < 0) return av_normalize_sf(av_normalize1_sf((SoftFloat){ b.mant + (a.mant >> (-t)), b.exp})); + else if (t < 32) return av_normalize_sf(av_normalize1_sf((SoftFloat){ a.mant + (b.mant >> t ), a.exp})); + else return a; +} + +static inline av_const SoftFloat av_sub_sf(SoftFloat a, SoftFloat b){ + return av_add_sf(a, (SoftFloat){ -b.mant, b.exp}); +} + +//FIXME log, exp, pow + +/** + * Converts a mantisse and exponent to a SoftFloat + * @returns a SoftFloat with value v * 2^frac_bits + */ +static inline av_const SoftFloat av_int2sf(int v, int frac_bits){ + int exp_offset = 0; + if(v == INT_MIN){ + exp_offset = 1; + v>>=1; + } + return av_normalize_sf(av_normalize1_sf((SoftFloat){v, ONE_BITS + 1 - frac_bits + exp_offset})); +} + +/** + * Rounding is to -inf. + */ +static inline av_const int av_sf2int(SoftFloat v, int frac_bits){ + v.exp += frac_bits - (ONE_BITS + 1); + if(v.exp >= 0) return v.mant << v.exp ; + else return v.mant >>(-v.exp); +} + +/** + * Rounding-to-nearest used. + */ +static av_always_inline SoftFloat av_sqrt_sf(SoftFloat val) +{ + int tabIndex, rem; + + if (val.mant == 0) + val.exp = MIN_EXP; + else if (val.mant < 0) + abort(); + else + { + tabIndex = (val.mant - 0x20000000) >> 20; + + rem = val.mant & 0xFFFFF; + val.mant = (int)(((int64_t)av_sqrttbl_sf[tabIndex] * (0x100000 - rem) + + (int64_t)av_sqrttbl_sf[tabIndex + 1] * rem + + 0x80000) >> 20); + val.mant = (int)(((int64_t)av_sqr_exp_multbl_sf[val.exp & 1] * val.mant + + 0x10000000) >> 29); + + if (val.mant < 0x40000000) + val.exp -= 2; + else + val.mant >>= 1; + + val.exp = (val.exp >> 1) + 1; + } + + return val; +} + +/** + * Rounding-to-nearest used. + */ +static av_unused void av_sincos_sf(int a, int *s, int *c) +{ + int idx, sign; + int sv, cv; + int st, ct; + + idx = a >> 26; + sign = (idx << 27) >> 31; + cv = av_costbl_1_sf[idx & 0xf]; + cv = (cv ^ sign) - sign; + + idx -= 8; + sign = (idx << 27) >> 31; + sv = av_costbl_1_sf[idx & 0xf]; + sv = (sv ^ sign) - sign; + + idx = a >> 21; + ct = av_costbl_2_sf[idx & 0x1f]; + st = av_sintbl_2_sf[idx & 0x1f]; + + idx = (int)(((int64_t)cv * ct - (int64_t)sv * st + 0x20000000) >> 30); + + sv = (int)(((int64_t)cv * st + (int64_t)sv * ct + 0x20000000) >> 30); + + cv = idx; + + idx = a >> 16; + ct = av_costbl_3_sf[idx & 0x1f]; + st = av_sintbl_3_sf[idx & 0x1f]; + + idx = (int)(((int64_t)cv * ct - (int64_t)sv * st + 0x20000000) >> 30); + + sv = (int)(((int64_t)cv * st + (int64_t)sv * ct + 0x20000000) >> 30); + cv = idx; + + idx = a >> 11; + + ct = (int)(((int64_t)av_costbl_4_sf[idx & 0x1f] * (0x800 - (a & 0x7ff)) + + (int64_t)av_costbl_4_sf[(idx & 0x1f)+1]*(a & 0x7ff) + + 0x400) >> 11); + st = (int)(((int64_t)av_sintbl_4_sf[idx & 0x1f] * (0x800 - (a & 0x7ff)) + + (int64_t)av_sintbl_4_sf[(idx & 0x1f) + 1] * (a & 0x7ff) + + 0x400) >> 11); + + *c = (int)(((int64_t)cv * ct + (int64_t)sv * st + 0x20000000) >> 30); + + *s = (int)(((int64_t)cv * st + (int64_t)sv * ct + 0x20000000) >> 30); +} + +#endif /* AVUTIL_SOFTFLOAT_H */ diff --git a/libavutil/softfloat_tables.h b/libavutil/softfloat_tables.h new file mode 100644 index 0000000000..461f2b221d --- /dev/null +++ b/libavutil/softfloat_tables.h @@ -0,0 +1,262 @@ +/* + * Copyright (c) 2012 + * MIPS Technologies, Inc., California. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * 3. Neither the name of the MIPS Technologies, Inc., nor the names of is + * contributors may be used to endorse or promote products derived from + * this software without specific prior written permission. + * + * THIS SOFTWARE IS PROVIDED BY THE MIPS TECHNOLOGIES, INC. ``AS IS'' AND + * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE + * ARE DISCLAIMED. IN NO EVENT SHALL THE MIPS TECHNOLOGIES, INC. BE LIABLE + * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT + * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY + * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF + * SUCH DAMAGE. + * + * Author: Stanislav Ocovaj (stanislav.ocovaj imgtec com) + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ +#ifndef AVUTIL_SOFTFLOAT_TABLES_H +#define AVUTIL_SOFTFLOAT_TABLES_H + +#include <stdint.h> + +static const int32_t av_sqrttbl_sf[512+1] = { /* sqrt(x), 0.5<=x<1 */ + 0x2d413ccd,0x2d4c8bb3,0x2d57d7c6,0x2d63210a, + 0x2d6e677f,0x2d79ab2a,0x2d84ec0b,0x2d902a23, + 0x2d9b6578,0x2da69e08,0x2db1d3d6,0x2dbd06e6, + 0x2dc83738,0x2dd364ce,0x2dde8fac,0x2de9b7d2, + 0x2df4dd43,0x2e000000,0x2e0b200c,0x2e163d68, + 0x2e215816,0x2e2c701a,0x2e378573,0x2e429824, + 0x2e4da830,0x2e58b598,0x2e63c05d,0x2e6ec883, + 0x2e79ce0a,0x2e84d0f5,0x2e8fd144,0x2e9acefb, + 0x2ea5ca1b,0x2eb0c2a7,0x2ebbb89e,0x2ec6ac04, + 0x2ed19cda,0x2edc8b23,0x2ee776df,0x2ef26012, + 0x2efd46bb,0x2f082add,0x2f130c7b,0x2f1deb95, + 0x2f28c82e,0x2f33a246,0x2f3e79e1,0x2f494eff, + 0x2f5421a3,0x2f5ef1ce,0x2f69bf81,0x2f748abe, + 0x2f7f5388,0x2f8a19e0,0x2f94ddc7,0x2f9f9f3e, + 0x2faa5e48,0x2fb51ae8,0x2fbfd51c,0x2fca8ce9, + 0x2fd5424e,0x2fdff54e,0x2feaa5eb,0x2ff55426, + 0x30000000,0x300aa97b,0x3015509a,0x301ff55c, + 0x302a97c5,0x303537d5,0x303fd58e,0x304a70f2, + 0x30550a01,0x305fa0be,0x306a352a,0x3074c747, + 0x307f5716,0x3089e499,0x30946fd2,0x309ef8c0, + 0x30a97f67,0x30b403c7,0x30be85e2,0x30c905bb, + 0x30d38351,0x30ddfea6,0x30e877bc,0x30f2ee96, + 0x30fd6332,0x3107d594,0x311245bc,0x311cb3ad, + 0x31271f67,0x313188ec,0x313bf03d,0x3146555c, + 0x3150b84a,0x315b1909,0x31657798,0x316fd3fc, + 0x317a2e34,0x31848642,0x318edc28,0x31992fe5, + 0x31a3817d,0x31add0f0,0x31b81e40,0x31c2696e, + 0x31ccb27b,0x31d6f969,0x31e13e38,0x31eb80eb, + 0x31f5c182,0x32000000,0x320a3c65,0x321476b1, + 0x321eaee8,0x3228e50a,0x32331917,0x323d4b13, + 0x32477afc,0x3251a8d6,0x325bd4a2,0x3265fe5f, + 0x32702611,0x327a4bb8,0x32846f55,0x328e90e9, + 0x3298b076,0x32a2cdfd,0x32ace97e,0x32b702fd, + 0x32c11a79,0x32cb2ff3,0x32d5436d,0x32df54e9, + 0x32e96466,0x32f371e8,0x32fd7d6d,0x330786f9, + 0x33118e8c,0x331b9426,0x332597cb,0x332f9979, + 0x33399933,0x334396fa,0x334d92cf,0x33578cb2, + 0x336184a6,0x336b7aab,0x33756ec3,0x337f60ed, + 0x3389512d,0x33933f83,0x339d2bef,0x33a71672, + 0x33b0ff10,0x33bae5c7,0x33c4ca99,0x33cead88, + 0x33d88e95,0x33e26dbf,0x33ec4b09,0x33f62673, + 0x34000000,0x3409d7af,0x3413ad82,0x341d817a, + 0x34275397,0x343123db,0x343af248,0x3444bedd, + 0x344e899d,0x34585288,0x3462199f,0x346bdee3, + 0x3475a254,0x347f63f5,0x348923c6,0x3492e1c9, + 0x349c9dfe,0x34a65865,0x34b01101,0x34b9c7d2, + 0x34c37cda,0x34cd3018,0x34d6e18f,0x34e0913f, + 0x34ea3f29,0x34f3eb4d,0x34fd95ae,0x35073e4c, + 0x3510e528,0x351a8a43,0x35242d9d,0x352dcf39, + 0x35376f16,0x35410d36,0x354aa99a,0x35544442, + 0x355ddd2f,0x35677463,0x357109df,0x357a9da2, + 0x35842fb0,0x358dc007,0x35974ea9,0x35a0db98, + 0x35aa66d3,0x35b3f05c,0x35bd7833,0x35c6fe5a, + 0x35d082d3,0x35da059c,0x35e386b7,0x35ed0626, + 0x35f683e8,0x36000000,0x36097a6e,0x3612f331, + 0x361c6a4d,0x3625dfc1,0x362f538f,0x3638c5b7, + 0x36423639,0x364ba518,0x36551252,0x365e7deb, + 0x3667e7e2,0x36715039,0x367ab6f0,0x36841c07, + 0x368d7f81,0x3696e15d,0x36a0419d,0x36a9a040, + 0x36b2fd49,0x36bc58b8,0x36c5b28e,0x36cf0acb, + 0x36d86170,0x36e1b680,0x36eb09f8,0x36f45bdc, + 0x36fdac2b,0x3706fae7,0x37104810,0x371993a7, + 0x3722ddad,0x372c2622,0x37356d08,0x373eb25f, + 0x3747f629,0x37513865,0x375a7914,0x3763b838, + 0x376cf5d0,0x377631e0,0x377f6c64,0x3788a561, + 0x3791dcd6,0x379b12c4,0x37a4472c,0x37ad7a0e, + 0x37b6ab6a,0x37bfdb44,0x37c90999,0x37d2366d, + 0x37db61be,0x37e48b8e,0x37edb3de,0x37f6daae, + 0x38000000,0x380923d3,0x3812462a,0x381b6703, + 0x38248660,0x382da442,0x3836c0aa,0x383fdb97, + 0x3848f50c,0x38520d09,0x385b238d,0x3864389b, + 0x386d4c33,0x38765e55,0x387f6f01,0x38887e3b, + 0x38918c00,0x389a9853,0x38a3a334,0x38acaca3, + 0x38b5b4a3,0x38bebb32,0x38c7c051,0x38d0c402, + 0x38d9c645,0x38e2c71b,0x38ebc685,0x38f4c482, + 0x38fdc114,0x3906bc3c,0x390fb5fa,0x3918ae4f, + 0x3921a53a,0x392a9abe,0x39338edb,0x393c8192, + 0x394572e2,0x394e62ce,0x39575155,0x39603e77, + 0x39692a36,0x39721494,0x397afd8f,0x3983e527, + 0x398ccb60,0x3995b039,0x399e93b2,0x39a775cc, + 0x39b05689,0x39b935e8,0x39c213e9,0x39caf08e, + 0x39d3cbd9,0x39dca5c7,0x39e57e5b,0x39ee5596, + 0x39f72b77,0x3a000000,0x3a08d331,0x3a11a50a, + 0x3a1a758d,0x3a2344ba,0x3a2c1291,0x3a34df13, + 0x3a3daa41,0x3a46741b,0x3a4f3ca3,0x3a5803d7, + 0x3a60c9ba,0x3a698e4b,0x3a72518b,0x3a7b137c, + 0x3a83d41d,0x3a8c936f,0x3a955173,0x3a9e0e29, + 0x3aa6c992,0x3aaf83ae,0x3ab83c7e,0x3ac0f403, + 0x3ac9aa3c,0x3ad25f2c,0x3adb12d1,0x3ae3c52d, + 0x3aec7642,0x3af5260e,0x3afdd492,0x3b0681d0, + 0x3b0f2dc6,0x3b17d878,0x3b2081e4,0x3b292a0c, + 0x3b31d0f0,0x3b3a7690,0x3b431aec,0x3b4bbe06, + 0x3b545fdf,0x3b5d0077,0x3b659fcd,0x3b6e3de4, + 0x3b76daba,0x3b7f7651,0x3b8810aa,0x3b90a9c4, + 0x3b9941a1,0x3ba1d842,0x3baa6da5,0x3bb301cd, + 0x3bbb94b9,0x3bc4266a,0x3bccb6e2,0x3bd5461f, + 0x3bddd423,0x3be660ee,0x3beeec81,0x3bf776dc, + 0x3c000000,0x3c0887ed,0x3c110ea4,0x3c199426, + 0x3c221872,0x3c2a9b8a,0x3c331d6e,0x3c3b9e1d, + 0x3c441d9a,0x3c4c9be5,0x3c5518fd,0x3c5d94e3, + 0x3c660f98,0x3c6e891d,0x3c770172,0x3c7f7898, + 0x3c87ee8e,0x3c906356,0x3c98d6ef,0x3ca1495b, + 0x3ca9ba9a,0x3cb22aac,0x3cba9992,0x3cc3074c, + 0x3ccb73dc,0x3cd3df41,0x3cdc497b,0x3ce4b28c, + 0x3ced1a73,0x3cf58132,0x3cfde6c8,0x3d064b37, + 0x3d0eae7f,0x3d17109f,0x3d1f719a,0x3d27d16e, + 0x3d30301d,0x3d388da8,0x3d40ea0d,0x3d49454f, + 0x3d519f6d,0x3d59f867,0x3d625040,0x3d6aa6f6, + 0x3d72fc8b,0x3d7b50fe,0x3d83a451,0x3d8bf683, + 0x3d944796,0x3d9c9788,0x3da4e65c,0x3dad3412, + 0x3db580a9,0x3dbdcc24,0x3dc61680,0x3dce5fc0, + 0x3dd6a7e4,0x3ddeeeed,0x3de734d9,0x3def79ab, + 0x3df7bd62,0x3e000000,0x3e084184,0x3e1081ee, + 0x3e18c140,0x3e20ff7a,0x3e293c9c,0x3e3178a7, + 0x3e39b39a,0x3e41ed77,0x3e4a263d,0x3e525def, + 0x3e5a948b,0x3e62ca12,0x3e6afe85,0x3e7331e4, + 0x3e7b642f,0x3e839567,0x3e8bc58c,0x3e93f49f, + 0x3e9c22a1,0x3ea44f91,0x3eac7b6f,0x3eb4a63e, + 0x3ebccffb,0x3ec4f8aa,0x3ecd2049,0x3ed546d9, + 0x3edd6c5a,0x3ee590cd,0x3eedb433,0x3ef5d68c, + 0x3efdf7d7,0x3f061816,0x3f0e3749,0x3f165570, + 0x3f1e728c,0x3f268e9d,0x3f2ea9a4,0x3f36c3a0, + 0x3f3edc93,0x3f46f47c,0x3f4f0b5d,0x3f572135, + 0x3f5f3606,0x3f6749cf,0x3f6f5c90,0x3f776e4a, + 0x3f7f7efe,0x3f878eab,0x3f8f9d53,0x3f97aaf6, + 0x3f9fb793,0x3fa7c32c,0x3fafcdc1,0x3fb7d752, + 0x3fbfdfe0,0x3fc7e76b,0x3fcfedf3,0x3fd7f378, + 0x3fdff7fc,0x3fe7fb7f,0x3feffe00,0x3ff7ff80, + 0x3fffffff, +}; + +static const int32_t av_sqr_exp_multbl_sf[2] = { + 0x20000000,0x2d413ccd, +}; + +static const int32_t av_costbl_1_sf[16] = { + 0x40000000,0x3ec52fa0,0x3b20d79e,0x3536cc52, + 0x2d413ccd,0x238e7673,0x187de2a7,0x0c7c5c1e, + 0x00000000,0xf383a3e3,0xe7821d5a,0xdc71898e, + 0xd2bec334,0xcac933af,0xc4df2863,0xc13ad061, +}; + +static const int32_t av_costbl_2_sf[32] = { + 0x40000000,0x3fffb10b,0x3ffec42d,0x3ffd3969, + 0x3ffb10c1,0x3ff84a3c,0x3ff4e5e0,0x3ff0e3b6, + 0x3fec43c7,0x3fe7061f,0x3fe12acb,0x3fdab1d9, + 0x3fd39b5a,0x3fcbe75e,0x3fc395f9,0x3fbaa740, + 0x3fb11b48,0x3fa6f228,0x3f9c2bfb,0x3f90c8da, + 0x3f84c8e2,0x3f782c30,0x3f6af2e3,0x3f5d1d1d, + 0x3f4eaafe,0x3f3f9cab,0x3f2ff24a,0x3f1fabff, + 0x3f0ec9f5,0x3efd4c54,0x3eeb3347,0x3ed87efc, +}; + +static const int32_t av_sintbl_2_sf[32] = { + 0x00000000,0x006487c4,0x00c90e90,0x012d936c, + 0x0192155f,0x01f69373,0x025b0caf,0x02bf801a, + 0x0323ecbe,0x038851a2,0x03ecadcf,0x0451004d, + 0x04b54825,0x0519845e,0x057db403,0x05e1d61b, + 0x0645e9af,0x06a9edc9,0x070de172,0x0771c3b3, + 0x07d59396,0x08395024,0x089cf867,0x09008b6a, + 0x09640837,0x09c76dd8,0x0a2abb59,0x0a8defc3, + 0x0af10a22,0x0b540982,0x0bb6ecef,0x0c19b374, +}; + +static const int32_t av_costbl_3_sf[32] = { + 0x40000000,0x3fffffec,0x3fffffb1,0x3fffff4e, + 0x3ffffec4,0x3ffffe13,0x3ffffd39,0x3ffffc39, + 0x3ffffb11,0x3ffff9c1,0x3ffff84a,0x3ffff6ac, + 0x3ffff4e6,0x3ffff2f8,0x3ffff0e3,0x3fffeea7, + 0x3fffec43,0x3fffe9b7,0x3fffe705,0x3fffe42a, + 0x3fffe128,0x3fffddff,0x3fffdaae,0x3fffd736, + 0x3fffd396,0x3fffcfcf,0x3fffcbe0,0x3fffc7ca, + 0x3fffc38c,0x3fffbf27,0x3fffba9b,0x3fffb5e7, +}; + +static const int32_t av_sintbl_3_sf[32] = { + 0x00000000,0x0003243f,0x0006487f,0x00096cbe, + 0x000c90fe,0x000fb53d,0x0012d97c,0x0015fdbb, + 0x001921fb,0x001c463a,0x001f6a79,0x00228eb8, + 0x0025b2f7,0x0028d736,0x002bfb74,0x002f1fb3, + 0x003243f1,0x00356830,0x00388c6e,0x003bb0ac, + 0x003ed4ea,0x0041f928,0x00451d66,0x004841a3, + 0x004b65e1,0x004e8a1e,0x0051ae5b,0x0054d297, + 0x0057f6d4,0x005b1b10,0x005e3f4c,0x00616388, +}; + +static const int32_t av_costbl_4_sf[33] = { + 0x40000000,0x40000000,0x40000000,0x40000000, + 0x40000000,0x40000000,0x3fffffff,0x3fffffff, + 0x3fffffff,0x3ffffffe,0x3ffffffe,0x3ffffffe, + 0x3ffffffd,0x3ffffffd,0x3ffffffc,0x3ffffffc, + 0x3ffffffb,0x3ffffffa,0x3ffffffa,0x3ffffff9, + 0x3ffffff8,0x3ffffff7,0x3ffffff7,0x3ffffff6, + 0x3ffffff5,0x3ffffff4,0x3ffffff3,0x3ffffff2, + 0x3ffffff1,0x3ffffff0,0x3fffffef,0x3fffffed, + 0x3fffffec, +}; + +static const int32_t av_sintbl_4_sf[33] = { + 0x00000000,0x00001922,0x00003244,0x00004b66, + 0x00006488,0x00007daa,0x000096cc,0x0000afee, + 0x0000c910,0x0000e232,0x0000fb54,0x00011476, + 0x00012d98,0x000146ba,0x00015fdc,0x000178fe, + 0x00019220,0x0001ab42,0x0001c464,0x0001dd86, + 0x0001f6a8,0x00020fca,0x000228ec,0x0002420e, + 0x00025b30,0x00027452,0x00028d74,0x0002a696, + 0x0002bfb7,0x0002d8d9,0x0002f1fb,0x00030b1d, + 0x0003243f, +}; +#endif /* AVUTIL_SOFTFLOAT_TABLES_H */ diff --git a/libavutil/stereo3d.c b/libavutil/stereo3d.c index 2dcfddf14b..50cd928de4 100644 --- a/libavutil/stereo3d.c +++ b/libavutil/stereo3d.c @@ -1,20 +1,20 @@ /* * Copyright (c) 2013 Vittorio Giovara <vittorio.giovara@gmail.com> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/stereo3d.h b/libavutil/stereo3d.h index b1910b1f87..1135dc9ddc 100644 --- a/libavutil/stereo3d.h +++ b/libavutil/stereo3d.h @@ -1,20 +1,20 @@ /* * Copyright (c) 2013 Vittorio Giovara <vittorio.giovara@gmail.com> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/tablegen.h b/libavutil/tablegen.h new file mode 100644 index 0000000000..02acdd61e8 --- /dev/null +++ b/libavutil/tablegen.h @@ -0,0 +1,55 @@ +/* + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +/** + * @file + * Compatibility libm for table generation files + */ + +#ifndef AVUTIL_TABLEGEN_H +#define AVUTIL_TABLEGEN_H + +#include <math.h> + +// we lack some functions on all host platforms, and we don't care about +// performance and/or strict ISO C semantics as it's performed at build time +static inline double ff_cbrt(double x) +{ + return x < 0 ? -pow(-x, 1.0 / 3.0) : pow(x, 1.0 / 3.0); +} +#define cbrt ff_cbrt + +static inline double ff_rint(double x) +{ + return x >= 0 ? floor(x + 0.5) : ceil(x - 0.5); +} +#define rint ff_rint + +static inline long long ff_llrint(double x) +{ + return rint(x); +} +#define llrint ff_llrint + +static inline long ff_lrint(double x) +{ + return rint(x); +} +#define lrint ff_lrint + +#endif /* AVUTIL_TABLEGEN_H */ diff --git a/libavutil/tea-test.c b/libavutil/tea-test.c new file mode 100644 index 0000000000..07ec56139a --- /dev/null +++ b/libavutil/tea-test.c @@ -0,0 +1,114 @@ +/* + * A 32-bit implementation of the TEA algorithm + * Copyright (c) 2015 Vesselin Bontchev + * + * Loosely based on the implementation of David Wheeler and Roger Needham, + * https://en.wikipedia.org/wiki/Tiny_Encryption_Algorithm#Reference_code + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "tea.c" + +#include <stdio.h> + +#define TEA_NUM_TESTS 4 + +// https://github.com/logandrews/TeaCrypt/blob/master/tea/tea_test.go +static const uint8_t tea_test_key[TEA_NUM_TESTS][16] = { + { 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00 + }, + { 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00 + }, + { 0x00, 0x11, 0x22, 0x33, 0x44, 0x55, 0x66, 0x77, + 0x88, 0x99, 0xAA, 0xBB, 0xCC, 0xDD, 0xEE, 0xFF + }, + { 0x00, 0x11, 0x22, 0x33, 0x44, 0x55, 0x66, 0x77, + 0x88, 0x99, 0xAA, 0xBB, 0xCC, 0xDD, 0xEE, 0xFF + } +}; + +static const uint8_t tea_test_pt[TEA_NUM_TESTS][8] = { + { 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00 }, + { 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08 }, + { 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08 }, + { 0x01, 0x23, 0x45, 0x67, 0x89, 0xAB, 0xCD, 0xEF } +}; + +static const uint8_t tea_test_ct[TEA_NUM_TESTS][8] = { + { 0x41, 0xEA, 0x3A, 0x0A, 0x94, 0xBA, 0xA9, 0x40 }, + { 0x6A, 0x2F, 0x9C, 0xF3, 0xFC, 0xCF, 0x3C, 0x55 }, + { 0xDE, 0xB1, 0xC0, 0xA2, 0x7E, 0x74, 0x5D, 0xB3 }, + { 0x12, 0x6C, 0x6B, 0x92, 0xC0, 0x65, 0x3A, 0x3E } +}; + +static void test_tea(AVTEA *ctx, uint8_t *dst, const uint8_t *src, + const uint8_t *ref, int len, uint8_t *iv, int dir, + const char *test) +{ + av_tea_crypt(ctx, dst, src, len, iv, dir); + if (memcmp(dst, ref, 8*len)) { + int i; + printf("%s failed\ngot ", test); + for (i = 0; i < 8*len; i++) + printf("%02x ", dst[i]); + printf("\nexpected "); + for (i = 0; i < 8*len; i++) + printf("%02x ", ref[i]); + printf("\n"); + exit(1); + } +} + +int main(void) +{ + AVTEA *ctx; + uint8_t buf[8], iv[8]; + int i; + static const uint8_t src[32] = "HelloWorldHelloWorldHelloWorld"; + uint8_t ct[32]; + uint8_t pl[32]; + + ctx = av_tea_alloc(); + if (!ctx) + return 1; + + for (i = 0; i < TEA_NUM_TESTS; i++) { + av_tea_init(ctx, tea_test_key[i], 64); + + test_tea(ctx, buf, tea_test_pt[i], tea_test_ct[i], 1, NULL, 0, "encryption"); + test_tea(ctx, buf, tea_test_ct[i], tea_test_pt[i], 1, NULL, 1, "decryption"); + + /* encrypt */ + memcpy(iv, "HALLO123", 8); + av_tea_crypt(ctx, ct, src, 4, iv, 0); + + /* decrypt into pl */ + memcpy(iv, "HALLO123", 8); + test_tea(ctx, pl, ct, src, 4, iv, 1, "CBC decryption"); + + memcpy(iv, "HALLO123", 8); + test_tea(ctx, ct, ct, src, 4, iv, 1, "CBC inplace decryption"); + } + + printf("Test encryption/decryption success.\n"); + av_free(ctx); + + return 0; +} diff --git a/libavutil/tea.c b/libavutil/tea.c new file mode 100644 index 0000000000..b138f8bea1 --- /dev/null +++ b/libavutil/tea.c @@ -0,0 +1,121 @@ +/* + * A 32-bit implementation of the TEA algorithm + * Copyright (c) 2015 Vesselin Bontchev + * + * Loosely based on the implementation of David Wheeler and Roger Needham, + * https://en.wikipedia.org/wiki/Tiny_Encryption_Algorithm#Reference_code + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "avutil.h" +#include "common.h" +#include "intreadwrite.h" +#include "tea.h" + +typedef struct AVTEA { + uint32_t key[16]; + int rounds; +} AVTEA; + +struct AVTEA *av_tea_alloc(void) +{ + return av_mallocz(sizeof(struct AVTEA)); +} + +const int av_tea_size = sizeof(AVTEA); + +void av_tea_init(AVTEA *ctx, const uint8_t key[16], int rounds) +{ + int i; + + for (i = 0; i < 4; i++) + ctx->key[i] = AV_RB32(key + (i << 2)); + + ctx->rounds = rounds; +} + +static void tea_crypt_ecb(AVTEA *ctx, uint8_t *dst, const uint8_t *src, + int decrypt, uint8_t *iv) +{ + uint32_t v0, v1; + int rounds = ctx->rounds; + uint32_t k0, k1, k2, k3; + k0 = ctx->key[0]; + k1 = ctx->key[1]; + k2 = ctx->key[2]; + k3 = ctx->key[3]; + + v0 = AV_RB32(src); + v1 = AV_RB32(src + 4); + + if (decrypt) { + int i; + uint32_t delta = 0x9E3779B9U, sum = delta * (rounds / 2); + + for (i = 0; i < rounds / 2; i++) { + v1 -= ((v0 << 4) + k2) ^ (v0 + sum) ^ ((v0 >> 5) + k3); + v0 -= ((v1 << 4) + k0) ^ (v1 + sum) ^ ((v1 >> 5) + k1); + sum -= delta; + } + if (iv) { + v0 ^= AV_RB32(iv); + v1 ^= AV_RB32(iv + 4); + memcpy(iv, src, 8); + } + } else { + int i; + uint32_t sum = 0, delta = 0x9E3779B9U; + + for (i = 0; i < rounds / 2; i++) { + sum += delta; + v0 += ((v1 << 4) + k0) ^ (v1 + sum) ^ ((v1 >> 5) + k1); + v1 += ((v0 << 4) + k2) ^ (v0 + sum) ^ ((v0 >> 5) + k3); + } + } + + AV_WB32(dst, v0); + AV_WB32(dst + 4, v1); +} + +void av_tea_crypt(AVTEA *ctx, uint8_t *dst, const uint8_t *src, int count, + uint8_t *iv, int decrypt) +{ + int i; + + if (decrypt) { + while (count--) { + tea_crypt_ecb(ctx, dst, src, decrypt, iv); + + src += 8; + dst += 8; + } + } else { + while (count--) { + if (iv) { + for (i = 0; i < 8; i++) + dst[i] = src[i] ^ iv[i]; + tea_crypt_ecb(ctx, dst, dst, decrypt, NULL); + memcpy(iv, dst, 8); + } else { + tea_crypt_ecb(ctx, dst, src, decrypt, NULL); + } + src += 8; + dst += 8; + } + } +} diff --git a/libavutil/tea.h b/libavutil/tea.h new file mode 100644 index 0000000000..dd929bdafd --- /dev/null +++ b/libavutil/tea.h @@ -0,0 +1,71 @@ +/* + * A 32-bit implementation of the TEA algorithm + * Copyright (c) 2015 Vesselin Bontchev + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#ifndef AVUTIL_TEA_H +#define AVUTIL_TEA_H + +#include <stdint.h> + +/** + * @file + * @brief Public header for libavutil TEA algorithm + * @defgroup lavu_tea TEA + * @ingroup lavu_crypto + * @{ + */ + +extern const int av_tea_size; + +struct AVTEA; + +/** + * Allocate an AVTEA context + * To free the struct: av_free(ptr) + */ +struct AVTEA *av_tea_alloc(void); + +/** + * Initialize an AVTEA context. + * + * @param ctx an AVTEA context + * @param key a key of 16 bytes used for encryption/decryption + * @param rounds the number of rounds in TEA (64 is the "standard") + */ +void av_tea_init(struct AVTEA *ctx, const uint8_t key[16], int rounds); + +/** + * Encrypt or decrypt a buffer using a previously initialized context. + * + * @param ctx an AVTEA context + * @param dst destination array, can be equal to src + * @param src source array, can be equal to dst + * @param count number of 8 byte blocks + * @param iv initialization vector for CBC mode, if NULL then ECB will be used + * @param decrypt 0 for encryption, 1 for decryption + */ +void av_tea_crypt(struct AVTEA *ctx, uint8_t *dst, const uint8_t *src, + int count, uint8_t *iv, int decrypt); + +/** + * @} + */ + +#endif /* AVUTIL_TEA_H */ diff --git a/libavutil/thread.h b/libavutil/thread.h index cf0fbdda5f..32ddf40365 100644 --- a/libavutil/thread.h +++ b/libavutil/thread.h @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -24,10 +24,110 @@ #include "config.h" -#if HAVE_PTHREADS || HAVE_W32THREADS +#if HAVE_PTHREADS || HAVE_W32THREADS || HAVE_OS2THREADS + +#define USE_ATOMICS 0 #if HAVE_PTHREADS #include <pthread.h> + +#if defined(ASSERT_LEVEL) && ASSERT_LEVEL > 1 + +#include "log.h" + +#define ASSERT_PTHREAD_NORET(func, ...) do { \ + int ret = func(__VA_ARGS__); \ + if (ret) { \ + av_log(NULL, AV_LOG_FATAL, AV_STRINGIFY(func) \ + " failed with error: %s\n", av_err2str(AVERROR(ret))); \ + abort(); \ + } \ +} while (0) + +#define ASSERT_PTHREAD(func, ...) do { \ + ASSERT_PTHREAD_NORET(func, __VA_ARGS__); \ + return 0; \ +} while (0) + +static inline int strict_pthread_join(pthread_t thread, void **value_ptr) +{ + ASSERT_PTHREAD(pthread_join, thread, value_ptr); +} + +static inline int strict_pthread_mutex_init(pthread_mutex_t *mutex, const pthread_mutexattr_t *attr) +{ + if (attr) { + ASSERT_PTHREAD_NORET(pthread_mutex_init, mutex, attr); + } else { + pthread_mutexattr_t local_attr; + ASSERT_PTHREAD_NORET(pthread_mutexattr_init, &local_attr); + ASSERT_PTHREAD_NORET(pthread_mutexattr_settype, &local_attr, PTHREAD_MUTEX_ERRORCHECK); + ASSERT_PTHREAD_NORET(pthread_mutex_init, mutex, &local_attr); + ASSERT_PTHREAD_NORET(pthread_mutexattr_destroy, &local_attr); + } + return 0; +} + +static inline int strict_pthread_mutex_destroy(pthread_mutex_t *mutex) +{ + ASSERT_PTHREAD(pthread_mutex_destroy, mutex); +} + +static inline int strict_pthread_mutex_lock(pthread_mutex_t *mutex) +{ + ASSERT_PTHREAD(pthread_mutex_lock, mutex); +} + +static inline int strict_pthread_mutex_unlock(pthread_mutex_t *mutex) +{ + ASSERT_PTHREAD(pthread_mutex_unlock, mutex); +} + +static inline int strict_pthread_cond_init(pthread_cond_t *cond, const pthread_condattr_t *attr) +{ + ASSERT_PTHREAD(pthread_cond_init, cond, attr); +} + +static inline int strict_pthread_cond_destroy(pthread_cond_t *cond) +{ + ASSERT_PTHREAD(pthread_cond_destroy, cond); +} + +static inline int strict_pthread_cond_signal(pthread_cond_t *cond) +{ + ASSERT_PTHREAD(pthread_cond_signal, cond); +} + +static inline int strict_pthread_cond_broadcast(pthread_cond_t *cond) +{ + ASSERT_PTHREAD(pthread_cond_broadcast, cond); +} + +static inline int strict_pthread_cond_wait(pthread_cond_t *cond, pthread_mutex_t *mutex) +{ + ASSERT_PTHREAD(pthread_cond_wait, cond, mutex); +} + +static inline int strict_pthread_once(pthread_once_t *once_control, void (*init_routine)(void)) +{ + ASSERT_PTHREAD(pthread_once, once_control, init_routine); +} + +#define pthread_join strict_pthread_join +#define pthread_mutex_init strict_pthread_mutex_init +#define pthread_mutex_destroy strict_pthread_mutex_destroy +#define pthread_mutex_lock strict_pthread_mutex_lock +#define pthread_mutex_unlock strict_pthread_mutex_unlock +#define pthread_cond_init strict_pthread_cond_init +#define pthread_cond_destroy strict_pthread_cond_destroy +#define pthread_cond_signal strict_pthread_cond_signal +#define pthread_cond_broadcast strict_pthread_cond_broadcast +#define pthread_cond_wait strict_pthread_cond_wait +#define pthread_once strict_pthread_once +#endif + +#elif HAVE_OS2THREADS +#include "compat/os2threads.h" #else #include "compat/w32pthreads.h" #endif @@ -46,6 +146,8 @@ #else +#define USE_ATOMICS 1 + #define AVMutex char #define ff_mutex_init(mutex, attr) (0) diff --git a/libavutil/threadmessage.c b/libavutil/threadmessage.c new file mode 100644 index 0000000000..7c5cd2463c --- /dev/null +++ b/libavutil/threadmessage.c @@ -0,0 +1,222 @@ +/* + * Copyright (c) 2014 Nicolas George + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public License + * as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public License + * along with FFmpeg; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "fifo.h" +#include "threadmessage.h" +#include "thread.h" + +struct AVThreadMessageQueue { +#if HAVE_THREADS + AVFifoBuffer *fifo; + pthread_mutex_t lock; + pthread_cond_t cond_recv; + pthread_cond_t cond_send; + int err_send; + int err_recv; + unsigned elsize; + void (*free_func)(void *msg); +#else + int dummy; +#endif +}; + +int av_thread_message_queue_alloc(AVThreadMessageQueue **mq, + unsigned nelem, + unsigned elsize) +{ +#if HAVE_THREADS + AVThreadMessageQueue *rmq; + int ret = 0; + + if (nelem > INT_MAX / elsize) + return AVERROR(EINVAL); + if (!(rmq = av_mallocz(sizeof(*rmq)))) + return AVERROR(ENOMEM); + if ((ret = pthread_mutex_init(&rmq->lock, NULL))) { + av_free(rmq); + return AVERROR(ret); + } + if ((ret = pthread_cond_init(&rmq->cond_recv, NULL))) { + pthread_mutex_destroy(&rmq->lock); + av_free(rmq); + return AVERROR(ret); + } + if ((ret = pthread_cond_init(&rmq->cond_send, NULL))) { + pthread_cond_destroy(&rmq->cond_recv); + pthread_mutex_destroy(&rmq->lock); + av_free(rmq); + return AVERROR(ret); + } + if (!(rmq->fifo = av_fifo_alloc(elsize * nelem))) { + pthread_cond_destroy(&rmq->cond_send); + pthread_cond_destroy(&rmq->cond_recv); + pthread_mutex_destroy(&rmq->lock); + av_free(rmq); + return AVERROR(ret); + } + rmq->elsize = elsize; + *mq = rmq; + return 0; +#else + *mq = NULL; + return AVERROR(ENOSYS); +#endif /* HAVE_THREADS */ +} + +void av_thread_message_queue_set_free_func(AVThreadMessageQueue *mq, + void (*free_func)(void *msg)) +{ +#if HAVE_THREADS + mq->free_func = free_func; +#endif +} + +void av_thread_message_queue_free(AVThreadMessageQueue **mq) +{ +#if HAVE_THREADS + if (*mq) { + av_thread_message_flush(*mq); + av_fifo_freep(&(*mq)->fifo); + pthread_cond_destroy(&(*mq)->cond_send); + pthread_cond_destroy(&(*mq)->cond_recv); + pthread_mutex_destroy(&(*mq)->lock); + av_freep(mq); + } +#endif +} + +#if HAVE_THREADS + +static int av_thread_message_queue_send_locked(AVThreadMessageQueue *mq, + void *msg, + unsigned flags) +{ + while (!mq->err_send && av_fifo_space(mq->fifo) < mq->elsize) { + if ((flags & AV_THREAD_MESSAGE_NONBLOCK)) + return AVERROR(EAGAIN); + pthread_cond_wait(&mq->cond_send, &mq->lock); + } + if (mq->err_send) + return mq->err_send; + av_fifo_generic_write(mq->fifo, msg, mq->elsize, NULL); + /* one message is sent, signal one receiver */ + pthread_cond_signal(&mq->cond_recv); + return 0; +} + +static int av_thread_message_queue_recv_locked(AVThreadMessageQueue *mq, + void *msg, + unsigned flags) +{ + while (!mq->err_recv && av_fifo_size(mq->fifo) < mq->elsize) { + if ((flags & AV_THREAD_MESSAGE_NONBLOCK)) + return AVERROR(EAGAIN); + pthread_cond_wait(&mq->cond_recv, &mq->lock); + } + if (av_fifo_size(mq->fifo) < mq->elsize) + return mq->err_recv; + av_fifo_generic_read(mq->fifo, msg, mq->elsize, NULL); + /* one message space appeared, signal one sender */ + pthread_cond_signal(&mq->cond_send); + return 0; +} + +#endif /* HAVE_THREADS */ + +int av_thread_message_queue_send(AVThreadMessageQueue *mq, + void *msg, + unsigned flags) +{ +#if HAVE_THREADS + int ret; + + pthread_mutex_lock(&mq->lock); + ret = av_thread_message_queue_send_locked(mq, msg, flags); + pthread_mutex_unlock(&mq->lock); + return ret; +#else + return AVERROR(ENOSYS); +#endif /* HAVE_THREADS */ +} + +int av_thread_message_queue_recv(AVThreadMessageQueue *mq, + void *msg, + unsigned flags) +{ +#if HAVE_THREADS + int ret; + + pthread_mutex_lock(&mq->lock); + ret = av_thread_message_queue_recv_locked(mq, msg, flags); + pthread_mutex_unlock(&mq->lock); + return ret; +#else + return AVERROR(ENOSYS); +#endif /* HAVE_THREADS */ +} + +void av_thread_message_queue_set_err_send(AVThreadMessageQueue *mq, + int err) +{ +#if HAVE_THREADS + pthread_mutex_lock(&mq->lock); + mq->err_send = err; + pthread_cond_broadcast(&mq->cond_send); + pthread_mutex_unlock(&mq->lock); +#endif /* HAVE_THREADS */ +} + +void av_thread_message_queue_set_err_recv(AVThreadMessageQueue *mq, + int err) +{ +#if HAVE_THREADS + pthread_mutex_lock(&mq->lock); + mq->err_recv = err; + pthread_cond_broadcast(&mq->cond_recv); + pthread_mutex_unlock(&mq->lock); +#endif /* HAVE_THREADS */ +} + +#if HAVE_THREADS +static void free_func_wrap(void *arg, void *msg, int size) +{ + AVThreadMessageQueue *mq = arg; + mq->free_func(msg); +} +#endif + +void av_thread_message_flush(AVThreadMessageQueue *mq) +{ +#if HAVE_THREADS + int used, off; + void *free_func = mq->free_func; + + pthread_mutex_lock(&mq->lock); + used = av_fifo_size(mq->fifo); + if (free_func) + for (off = 0; off < used; off += mq->elsize) + av_fifo_generic_peek_at(mq->fifo, mq, off, mq->elsize, free_func_wrap); + av_fifo_drain(mq->fifo, used); + /* only the senders need to be notified since the queue is empty and there + * is nothing to read */ + pthread_cond_broadcast(&mq->cond_send); + pthread_mutex_unlock(&mq->lock); +#endif /* HAVE_THREADS */ +} diff --git a/libavutil/threadmessage.h b/libavutil/threadmessage.h new file mode 100644 index 0000000000..e256cae9db --- /dev/null +++ b/libavutil/threadmessage.h @@ -0,0 +1,107 @@ +/* + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public License + * as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public License + * along with FFmpeg; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#ifndef AVUTIL_THREADMESSAGE_H +#define AVUTIL_THREADMESSAGE_H + +typedef struct AVThreadMessageQueue AVThreadMessageQueue; + +typedef enum AVThreadMessageFlags { + + /** + * Perform non-blocking operation. + * If this flag is set, send and recv operations are non-blocking and + * return AVERROR(EAGAIN) immediately if they can not proceed. + */ + AV_THREAD_MESSAGE_NONBLOCK = 1, + +} AVThreadMessageFlags; + +/** + * Allocate a new message queue. + * + * @param mq pointer to the message queue + * @param nelem maximum number of elements in the queue + * @param elsize size of each element in the queue + * @return >=0 for success; <0 for error, in particular AVERROR(ENOSYS) if + * lavu was built without thread support + */ +int av_thread_message_queue_alloc(AVThreadMessageQueue **mq, + unsigned nelem, + unsigned elsize); + +/** + * Free a message queue. + * + * The message queue must no longer be in use by another thread. + */ +void av_thread_message_queue_free(AVThreadMessageQueue **mq); + +/** + * Send a message on the queue. + */ +int av_thread_message_queue_send(AVThreadMessageQueue *mq, + void *msg, + unsigned flags); + +/** + * Receive a message from the queue. + */ +int av_thread_message_queue_recv(AVThreadMessageQueue *mq, + void *msg, + unsigned flags); + +/** + * Set the sending error code. + * + * If the error code is set to non-zero, av_thread_message_queue_recv() will + * return it immediately when there are no longer available messages. + * Conventional values, such as AVERROR_EOF or AVERROR(EAGAIN), can be used + * to cause the receiving thread to stop or suspend its operation. + */ +void av_thread_message_queue_set_err_send(AVThreadMessageQueue *mq, + int err); + +/** + * Set the receiving error code. + * + * If the error code is set to non-zero, av_thread_message_queue_send() will + * return it immediately. Conventional values, such as AVERROR_EOF or + * AVERROR(EAGAIN), can be used to cause the sending thread to stop or + * suspend its operation. + */ +void av_thread_message_queue_set_err_recv(AVThreadMessageQueue *mq, + int err); + +/** + * Set the optional free message callback function which will be called if an + * operation is removing messages from the queue. + */ +void av_thread_message_queue_set_free_func(AVThreadMessageQueue *mq, + void (*free_func)(void *msg)); + +/** + * Flush the message queue + * + * This function is mostly equivalent to reading and free-ing every message + * except that it will be done in a single operation (no lock/unlock between + * reads). + */ +void av_thread_message_flush(AVThreadMessageQueue *mq); + +#endif /* AVUTIL_THREADMESSAGE_H */ diff --git a/libavutil/time.c b/libavutil/time.c index 7a3604e651..dbaee0264c 100644 --- a/libavutil/time.c +++ b/libavutil/time.c @@ -1,18 +1,20 @@ /* - * This file is part of Libav. + * Copyright (c) 2000-2003 Fabrice Bellard * - * Libav is free software; you can redistribute it and/or + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -21,9 +23,6 @@ #include <stddef.h> #include <stdint.h> #include <time.h> -#if HAVE_CLOCK_GETTIME -#include <time.h> -#endif #if HAVE_GETTIMEOFDAY #include <sys/time.h> #endif @@ -56,7 +55,7 @@ int64_t av_gettime(void) int64_t av_gettime_relative(void) { -#if HAVE_CLOCK_GETTIME +#if HAVE_CLOCK_GETTIME && defined(CLOCK_MONOTONIC) struct timespec ts; clock_gettime(CLOCK_MONOTONIC, &ts); return (int64_t)ts.tv_sec * 1000000 + ts.tv_nsec / 1000; @@ -65,6 +64,15 @@ int64_t av_gettime_relative(void) #endif } +int av_gettime_relative_is_monotonic(void) +{ +#if HAVE_CLOCK_GETTIME && defined(CLOCK_MONOTONIC) + return 1; +#else + return 0; +#endif +} + int av_usleep(unsigned usec) { #if HAVE_NANOSLEEP diff --git a/libavutil/time.h b/libavutil/time.h index 463bcfcdcf..dc169b064a 100644 --- a/libavutil/time.h +++ b/libavutil/time.h @@ -1,18 +1,20 @@ /* - * This file is part of Libav. + * Copyright (c) 2000-2003 Fabrice Bellard * - * Libav is free software; you can redistribute it and/or + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -36,6 +38,12 @@ int64_t av_gettime(void); int64_t av_gettime_relative(void); /** + * Indicates with a boolean result if the av_gettime_relative() time source + * is monotonic. + */ +int av_gettime_relative_is_monotonic(void); + +/** * Sleep for a period of time. Although the duration is expressed in * microseconds, the actual delay may be rounded to the precision of the * system timer. diff --git a/libavutil/time_internal.h b/libavutil/time_internal.h index 829fefb007..612a75a041 100644 --- a/libavutil/time_internal.h +++ b/libavutil/time_internal.h @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -22,7 +22,7 @@ #include <time.h> #include "config.h" -#if !HAVE_GMTIME_R +#if !HAVE_GMTIME_R && !defined(gmtime_r) static inline struct tm *gmtime_r(const time_t* clock, struct tm *result) { struct tm *ptr = gmtime(clock); @@ -33,7 +33,7 @@ static inline struct tm *gmtime_r(const time_t* clock, struct tm *result) } #endif -#if !HAVE_LOCALTIME_R +#if !HAVE_LOCALTIME_R && !defined(localtime_r) static inline struct tm *localtime_r(const time_t* clock, struct tm *result) { struct tm *ptr = localtime(clock); diff --git a/libavutil/timecode.c b/libavutil/timecode.c new file mode 100644 index 0000000000..fa92df1ef9 --- /dev/null +++ b/libavutil/timecode.c @@ -0,0 +1,219 @@ +/* + * Copyright (c) 2006 Smartjog S.A.S, Baptiste Coudurier <baptiste.coudurier@gmail.com> + * Copyright (c) 2011-2012 Smartjog S.A.S, Clément Bœsch <clement.boesch@smartjog.com> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +/** + * @file + * Timecode helpers + * @see https://en.wikipedia.org/wiki/SMPTE_time_code + * @see http://www.dropframetimecode.org + */ + +#include <stdio.h> +#include "timecode.h" +#include "log.h" +#include "error.h" + +int av_timecode_adjust_ntsc_framenum2(int framenum, int fps) +{ + /* only works for NTSC 29.97 and 59.94 */ + int drop_frames = 0; + int d, m, frames_per_10mins; + + if (fps == 30) { + drop_frames = 2; + frames_per_10mins = 17982; + } else if (fps == 60) { + drop_frames = 4; + frames_per_10mins = 35964; + } else + return framenum; + + d = framenum / frames_per_10mins; + m = framenum % frames_per_10mins; + + return framenum + 9 * drop_frames * d + drop_frames * ((m - drop_frames) / (frames_per_10mins / 10)); +} + +uint32_t av_timecode_get_smpte_from_framenum(const AVTimecode *tc, int framenum) +{ + unsigned fps = tc->fps; + int drop = !!(tc->flags & AV_TIMECODE_FLAG_DROPFRAME); + int hh, mm, ss, ff; + + framenum += tc->start; + if (drop) + framenum = av_timecode_adjust_ntsc_framenum2(framenum, tc->fps); + ff = framenum % fps; + ss = framenum / fps % 60; + mm = framenum / (fps*60) % 60; + hh = framenum / (fps*3600) % 24; + return 0 << 31 | // color frame flag (0: unsync mode, 1: sync mode) + drop << 30 | // drop frame flag (0: non drop, 1: drop) + (ff / 10) << 28 | // tens of frames + (ff % 10) << 24 | // units of frames + 0 << 23 | // PC (NTSC) or BGF0 (PAL) + (ss / 10) << 20 | // tens of seconds + (ss % 10) << 16 | // units of seconds + 0 << 15 | // BGF0 (NTSC) or BGF2 (PAL) + (mm / 10) << 12 | // tens of minutes + (mm % 10) << 8 | // units of minutes + 0 << 7 | // BGF2 (NTSC) or PC (PAL) + 0 << 6 | // BGF1 + (hh / 10) << 4 | // tens of hours + (hh % 10); // units of hours +} + +char *av_timecode_make_string(const AVTimecode *tc, char *buf, int framenum) +{ + int fps = tc->fps; + int drop = tc->flags & AV_TIMECODE_FLAG_DROPFRAME; + int hh, mm, ss, ff, neg = 0; + + framenum += tc->start; + if (drop) + framenum = av_timecode_adjust_ntsc_framenum2(framenum, fps); + if (framenum < 0) { + framenum = -framenum; + neg = tc->flags & AV_TIMECODE_FLAG_ALLOWNEGATIVE; + } + ff = framenum % fps; + ss = framenum / fps % 60; + mm = framenum / (fps*60) % 60; + hh = framenum / (fps*3600); + if (tc->flags & AV_TIMECODE_FLAG_24HOURSMAX) + hh = hh % 24; + snprintf(buf, AV_TIMECODE_STR_SIZE, "%s%02d:%02d:%02d%c%02d", + neg ? "-" : "", + hh, mm, ss, drop ? ';' : ':', ff); + return buf; +} + +static unsigned bcd2uint(uint8_t bcd) +{ + unsigned low = bcd & 0xf; + unsigned high = bcd >> 4; + if (low > 9 || high > 9) + return 0; + return low + 10*high; +} + +char *av_timecode_make_smpte_tc_string(char *buf, uint32_t tcsmpte, int prevent_df) +{ + unsigned hh = bcd2uint(tcsmpte & 0x3f); // 6-bit hours + unsigned mm = bcd2uint(tcsmpte>>8 & 0x7f); // 7-bit minutes + unsigned ss = bcd2uint(tcsmpte>>16 & 0x7f); // 7-bit seconds + unsigned ff = bcd2uint(tcsmpte>>24 & 0x3f); // 6-bit frames + unsigned drop = tcsmpte & 1<<30 && !prevent_df; // 1-bit drop if not arbitrary bit + snprintf(buf, AV_TIMECODE_STR_SIZE, "%02u:%02u:%02u%c%02u", + hh, mm, ss, drop ? ';' : ':', ff); + return buf; +} + +char *av_timecode_make_mpeg_tc_string(char *buf, uint32_t tc25bit) +{ + snprintf(buf, AV_TIMECODE_STR_SIZE, "%02u:%02u:%02u%c%02u", + tc25bit>>19 & 0x1f, // 5-bit hours + tc25bit>>13 & 0x3f, // 6-bit minutes + tc25bit>>6 & 0x3f, // 6-bit seconds + tc25bit & 1<<24 ? ';' : ':', // 1-bit drop flag + tc25bit & 0x3f); // 6-bit frames + return buf; +} + +static int check_fps(int fps) +{ + int i; + static const int supported_fps[] = { + 24, 25, 30, 48, 50, 60, 100, 120, 150, + }; + + for (i = 0; i < FF_ARRAY_ELEMS(supported_fps); i++) + if (fps == supported_fps[i]) + return 0; + return -1; +} + +static int check_timecode(void *log_ctx, AVTimecode *tc) +{ + if ((int)tc->fps <= 0) { + av_log(log_ctx, AV_LOG_ERROR, "Timecode frame rate must be specified\n"); + return AVERROR(EINVAL); + } + if ((tc->flags & AV_TIMECODE_FLAG_DROPFRAME) && tc->fps != 30 && tc->fps != 60) { + av_log(log_ctx, AV_LOG_ERROR, "Drop frame is only allowed with 30000/1001 or 60000/1001 FPS\n"); + return AVERROR(EINVAL); + } + if (check_fps(tc->fps) < 0) { + av_log(log_ctx, AV_LOG_WARNING, "Using non-standard frame rate %d/%d\n", + tc->rate.num, tc->rate.den); + } + return 0; +} + +static int fps_from_frame_rate(AVRational rate) +{ + if (!rate.den || !rate.num) + return -1; + return (rate.num + rate.den/2) / rate.den; +} + +int av_timecode_check_frame_rate(AVRational rate) +{ + return check_fps(fps_from_frame_rate(rate)); +} + +int av_timecode_init(AVTimecode *tc, AVRational rate, int flags, int frame_start, void *log_ctx) +{ + memset(tc, 0, sizeof(*tc)); + tc->start = frame_start; + tc->flags = flags; + tc->rate = rate; + tc->fps = fps_from_frame_rate(rate); + return check_timecode(log_ctx, tc); +} + +int av_timecode_init_from_string(AVTimecode *tc, AVRational rate, const char *str, void *log_ctx) +{ + char c; + int hh, mm, ss, ff, ret; + + if (sscanf(str, "%d:%d:%d%c%d", &hh, &mm, &ss, &c, &ff) != 5) { + av_log(log_ctx, AV_LOG_ERROR, "Unable to parse timecode, " + "syntax: hh:mm:ss[:;.]ff\n"); + return AVERROR_INVALIDDATA; + } + + memset(tc, 0, sizeof(*tc)); + tc->flags = c != ':' ? AV_TIMECODE_FLAG_DROPFRAME : 0; // drop if ';', '.', ... + tc->rate = rate; + tc->fps = fps_from_frame_rate(rate); + + ret = check_timecode(log_ctx, tc); + if (ret < 0) + return ret; + + tc->start = (hh*3600 + mm*60 + ss) * tc->fps + ff; + if (tc->flags & AV_TIMECODE_FLAG_DROPFRAME) { /* adjust frame number */ + int tmins = 60*hh + mm; + tc->start -= 2 * (tmins - tmins/10); + } + return 0; +} diff --git a/libavutil/timecode.h b/libavutil/timecode.h new file mode 100644 index 0000000000..56e3975fd8 --- /dev/null +++ b/libavutil/timecode.h @@ -0,0 +1,140 @@ +/* + * Copyright (c) 2006 Smartjog S.A.S, Baptiste Coudurier <baptiste.coudurier@gmail.com> + * Copyright (c) 2011-2012 Smartjog S.A.S, Clément Bœsch <clement.boesch@smartjog.com> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +/** + * @file + * Timecode helpers header + */ + +#ifndef AVUTIL_TIMECODE_H +#define AVUTIL_TIMECODE_H + +#include <stdint.h> +#include "rational.h" + +#define AV_TIMECODE_STR_SIZE 16 + +enum AVTimecodeFlag { + AV_TIMECODE_FLAG_DROPFRAME = 1<<0, ///< timecode is drop frame + AV_TIMECODE_FLAG_24HOURSMAX = 1<<1, ///< timecode wraps after 24 hours + AV_TIMECODE_FLAG_ALLOWNEGATIVE = 1<<2, ///< negative time values are allowed +}; + +typedef struct { + int start; ///< timecode frame start (first base frame number) + uint32_t flags; ///< flags such as drop frame, +24 hours support, ... + AVRational rate; ///< frame rate in rational form + unsigned fps; ///< frame per second; must be consistent with the rate field +} AVTimecode; + +/** + * Adjust frame number for NTSC drop frame time code. + * + * @param framenum frame number to adjust + * @param fps frame per second, 30 or 60 + * @return adjusted frame number + * @warning adjustment is only valid in NTSC 29.97 and 59.94 + */ +int av_timecode_adjust_ntsc_framenum2(int framenum, int fps); + +/** + * Convert frame number to SMPTE 12M binary representation. + * + * @param tc timecode data correctly initialized + * @param framenum frame number + * @return the SMPTE binary representation + * + * @note Frame number adjustment is automatically done in case of drop timecode, + * you do NOT have to call av_timecode_adjust_ntsc_framenum2(). + * @note The frame number is relative to tc->start. + * @note Color frame (CF), binary group flags (BGF) and biphase mark polarity + * correction (PC) bits are set to zero. + */ +uint32_t av_timecode_get_smpte_from_framenum(const AVTimecode *tc, int framenum); + +/** + * Load timecode string in buf. + * + * @param buf destination buffer, must be at least AV_TIMECODE_STR_SIZE long + * @param tc timecode data correctly initialized + * @param framenum frame number + * @return the buf parameter + * + * @note Timecode representation can be a negative timecode and have more than + * 24 hours, but will only be honored if the flags are correctly set. + * @note The frame number is relative to tc->start. + */ +char *av_timecode_make_string(const AVTimecode *tc, char *buf, int framenum); + +/** + * Get the timecode string from the SMPTE timecode format. + * + * @param buf destination buffer, must be at least AV_TIMECODE_STR_SIZE long + * @param tcsmpte the 32-bit SMPTE timecode + * @param prevent_df prevent the use of a drop flag when it is known the DF bit + * is arbitrary + * @return the buf parameter + */ +char *av_timecode_make_smpte_tc_string(char *buf, uint32_t tcsmpte, int prevent_df); + +/** + * Get the timecode string from the 25-bit timecode format (MPEG GOP format). + * + * @param buf destination buffer, must be at least AV_TIMECODE_STR_SIZE long + * @param tc25bit the 25-bits timecode + * @return the buf parameter + */ +char *av_timecode_make_mpeg_tc_string(char *buf, uint32_t tc25bit); + +/** + * Init a timecode struct with the passed parameters. + * + * @param log_ctx a pointer to an arbitrary struct of which the first field + * is a pointer to an AVClass struct (used for av_log) + * @param tc pointer to an allocated AVTimecode + * @param rate frame rate in rational form + * @param flags miscellaneous flags such as drop frame, +24 hours, ... + * (see AVTimecodeFlag) + * @param frame_start the first frame number + * @return 0 on success, AVERROR otherwise + */ +int av_timecode_init(AVTimecode *tc, AVRational rate, int flags, int frame_start, void *log_ctx); + +/** + * Parse timecode representation (hh:mm:ss[:;.]ff). + * + * @param log_ctx a pointer to an arbitrary struct of which the first field is a + * pointer to an AVClass struct (used for av_log). + * @param tc pointer to an allocated AVTimecode + * @param rate frame rate in rational form + * @param str timecode string which will determine the frame start + * @return 0 on success, AVERROR otherwise + */ +int av_timecode_init_from_string(AVTimecode *tc, AVRational rate, const char *str, void *log_ctx); + +/** + * Check if the timecode feature is available for the given frame rate + * + * @return 0 if supported, <0 otherwise + */ +int av_timecode_check_frame_rate(AVRational rate); + +#endif /* AVUTIL_TIMECODE_H */ diff --git a/libavutil/timer.h b/libavutil/timer.h index cc4c5a5901..ed3b047870 100644 --- a/libavutil/timer.h +++ b/libavutil/timer.h @@ -1,20 +1,20 @@ /* * copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -42,8 +42,6 @@ # include "aarch64/timer.h" #elif ARCH_ARM # include "arm/timer.h" -#elif ARCH_BFIN -# include "bfin/timer.h" #elif ARCH_PPC # include "ppc/timer.h" #elif ARCH_X86 @@ -58,6 +56,10 @@ # endif #endif +#ifndef FF_TIMER_UNITS +# define FF_TIMER_UNITS "UNITS" +#endif + #ifdef AV_READ_TIME #define START_TIMER \ uint64_t tend; \ @@ -69,6 +71,8 @@ static uint64_t tsum = 0; \ static int tcount = 0; \ static int tskip_count = 0; \ + static int thistogram[32] = {0}; \ + thistogram[av_log2(tend - tstart)]++; \ if (tcount < 2 || \ tend - tstart < 8 * tsum / tcount || \ tend - tstart < 2000) { \ @@ -77,9 +81,13 @@ } else \ tskip_count++; \ if (((tcount + tskip_count) & (tcount + tskip_count - 1)) == 0) { \ + int i; \ av_log(NULL, AV_LOG_ERROR, \ - "%"PRIu64" UNITS in %s, %d runs, %d skips\n", \ + "%7"PRIu64" " FF_TIMER_UNITS " in %s,%8d runs,%7d skips", \ tsum * 10 / tcount, id, tcount, tskip_count); \ + for (i = 0; i < 32; i++) \ + av_log(NULL, AV_LOG_VERBOSE, " %2d", av_log2(2*thistogram[i]));\ + av_log(NULL, AV_LOG_ERROR, "\n"); \ } \ } #else diff --git a/libavutil/timestamp.h b/libavutil/timestamp.h new file mode 100644 index 0000000000..f010a7ee38 --- /dev/null +++ b/libavutil/timestamp.h @@ -0,0 +1,78 @@ +/* + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +/** + * @file + * timestamp utils, mostly useful for debugging/logging purposes + */ + +#ifndef AVUTIL_TIMESTAMP_H +#define AVUTIL_TIMESTAMP_H + +#include "common.h" + +#if defined(__cplusplus) && !defined(__STDC_FORMAT_MACROS) && !defined(PRId64) +#error missing -D__STDC_FORMAT_MACROS / #define __STDC_FORMAT_MACROS +#endif + +#define AV_TS_MAX_STRING_SIZE 32 + +/** + * Fill the provided buffer with a string containing a timestamp + * representation. + * + * @param buf a buffer with size in bytes of at least AV_TS_MAX_STRING_SIZE + * @param ts the timestamp to represent + * @return the buffer in input + */ +static inline char *av_ts_make_string(char *buf, int64_t ts) +{ + if (ts == AV_NOPTS_VALUE) snprintf(buf, AV_TS_MAX_STRING_SIZE, "NOPTS"); + else snprintf(buf, AV_TS_MAX_STRING_SIZE, "%"PRId64, ts); + return buf; +} + +/** + * Convenience macro, the return value should be used only directly in + * function arguments but never stand-alone. + */ +#define av_ts2str(ts) av_ts_make_string((char[AV_TS_MAX_STRING_SIZE]){0}, ts) + +/** + * Fill the provided buffer with a string containing a timestamp time + * representation. + * + * @param buf a buffer with size in bytes of at least AV_TS_MAX_STRING_SIZE + * @param ts the timestamp to represent + * @param tb the timebase of the timestamp + * @return the buffer in input + */ +static inline char *av_ts_make_time_string(char *buf, int64_t ts, AVRational *tb) +{ + if (ts == AV_NOPTS_VALUE) snprintf(buf, AV_TS_MAX_STRING_SIZE, "NOPTS"); + else snprintf(buf, AV_TS_MAX_STRING_SIZE, "%.6g", av_q2d(*tb) * ts); + return buf; +} + +/** + * Convenience macro, the return value should be used only directly in + * function arguments but never stand-alone. + */ +#define av_ts2timestr(ts, tb) av_ts_make_time_string((char[AV_TS_MAX_STRING_SIZE]){0}, ts, tb) + +#endif /* AVUTIL_TIMESTAMP_H */ diff --git a/libavutil/tomi/intreadwrite.h b/libavutil/tomi/intreadwrite.h index baf592a7e1..7dec4158d3 100644 --- a/libavutil/tomi/intreadwrite.h +++ b/libavutil/tomi/intreadwrite.h @@ -1,20 +1,20 @@ /* * Copyright (c) 2010 Mans Rullgard <mans@mansr.com> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/tree-test.c b/libavutil/tree-test.c index f93e53409d..d1df0b6c87 100644 --- a/libavutil/tree-test.c +++ b/libavutil/tree-test.c @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -54,34 +54,32 @@ static void print(AVTreeNode *t, int depth) av_log(NULL, AV_LOG_ERROR, "NULL\n"); } -static int cmp(void *a, const void *b) +static int cmp(const void *a, const void *b) { - return (uint8_t *) a - (const uint8_t *) b; + return (const uint8_t *) a - (const uint8_t *) b; } -int main(void) +int main(int argc, char **argv) { int i; + void *k; AVTreeNode *root = NULL, *node = NULL; AVLFG prng; + int log_level = argc <= 1 ? AV_LOG_INFO : atoi(argv[1]); + + av_log_set_level(log_level); av_lfg_init(&prng, 1); for (i = 0; i < 10000; i++) { - AVTreeNode *node2 = NULL; intptr_t j = av_lfg_get(&prng) % 86294; - void *ret, *jj = (void *)(j + 1); - - while (ret = av_tree_find(root, jj, cmp, NULL)) { - j = av_lfg_get(&prng) % 86294; - jj = (void *)(j + 1); - } if (check(root) > 999) { av_log(NULL, AV_LOG_ERROR, "FATAL error %d\n", i); print(root, 0); return 1; } + av_log(NULL, AV_LOG_DEBUG, "inserting %4d\n", (int)j); if (!node) node = av_tree_node_alloc(); @@ -89,20 +87,20 @@ int main(void) av_log(NULL, AV_LOG_ERROR, "Memory allocation failure.\n"); return 1; } - av_tree_insert(&root, jj, cmp, &node); - - while (ret = av_tree_find(root, jj, cmp, NULL)) { - j = av_lfg_get(&prng) % 86294; - jj = (void *)(j + 1); + av_tree_insert(&root, (void *)(j + 1), cmp, &node); + + j = av_lfg_get(&prng) % 86294; + { + AVTreeNode *node2 = NULL; + av_log(NULL, AV_LOG_DEBUG, "removing %4d\n", (int)j); + av_tree_insert(&root, (void *)(j + 1), cmp, &node2); + k = av_tree_find(root, (void *)(j + 1), cmp, NULL); + if (k) + av_log(NULL, AV_LOG_ERROR, "removal failure %d\n", i); + av_free(node2); } - - ret = av_tree_insert(&root, jj, cmp, &node2); - if (ret != jj) - av_tree_destroy(node2); - ret = av_tree_find(root, jj, cmp, NULL); - if (ret) - av_log(NULL, AV_LOG_ERROR, "removal failure %d\n", i); } + av_free(node); av_tree_destroy(root); diff --git a/libavutil/tree.c b/libavutil/tree.c index a65d728f86..7b57b2d39a 100644 --- a/libavutil/tree.c +++ b/libavutil/tree.c @@ -1,24 +1,25 @@ /* * copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ #include "error.h" +#include "log.h" #include "mem.h" #include "tree.h" @@ -28,13 +29,15 @@ typedef struct AVTreeNode { int state; } AVTreeNode; +const int av_tree_node_size = sizeof(AVTreeNode); + struct AVTreeNode *av_tree_node_alloc(void) { return av_mallocz(sizeof(struct AVTreeNode)); } void *av_tree_find(const AVTreeNode *t, void *key, - int (*cmp)(void *key, const void *b), void *next[2]) + int (*cmp)(const void *key, const void *b), void *next[2]) { if (t) { unsigned int v = cmp(key, t->elem); @@ -54,7 +57,7 @@ void *av_tree_find(const AVTreeNode *t, void *key, } void *av_tree_insert(AVTreeNode **tp, void *key, - int (*cmp)(void *key, const void *b), AVTreeNode **next) + int (*cmp)(const void *key, const void *b), AVTreeNode **next) { AVTreeNode *t = *tp; if (t) { diff --git a/libavutil/tree.h b/libavutil/tree.h index e2f191cdd7..9a9e11b92c 100644 --- a/libavutil/tree.h +++ b/libavutil/tree.h @@ -1,20 +1,20 @@ /* * copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -43,6 +43,7 @@ struct AVTreeNode; +extern const int av_tree_node_size; /** * Allocate an AVTreeNode. @@ -55,26 +56,34 @@ struct AVTreeNode *av_tree_node_alloc(void); * @param next If next is not NULL, then next[0] will contain the previous * element and next[1] the next element. If either does not exist, * then the corresponding entry in next is unchanged. + * @param cmp compare function used to compare elements in the tree, + * API identical to that of Standard C's qsort + * It is guranteed that the first and only the first argument to cmp() + * will be the key parameter to av_tree_find(), thus it could if the + * user wants, be a different type (like an opaque context). * @return An element with cmp(key, elem) == 0 or NULL if no such element * exists in the tree. */ void *av_tree_find(const struct AVTreeNode *root, void *key, - int (*cmp)(void *key, const void *b), void *next[2]); + int (*cmp)(const void *key, const void *b), void *next[2]); /** * Insert or remove an element. + * * If *next is NULL, then the supplied element will be removed if it exists. - * If *next is not NULL, then the supplied element will be inserted, unless + * If *next is non-NULL, then the supplied element will be inserted, unless * it already exists in the tree. + * * @param rootp A pointer to a pointer to the root node of the tree; note that * the root node can change during insertions, this is required * to keep the tree balanced. + * @param key pointer to the element key to insert in the tree * @param next Used to allocate and free AVTreeNodes. For insertion the user * must set it to an allocated and zeroed object of at least * av_tree_node_size bytes size. av_tree_insert() will set it to * NULL if it has been consumed. * For deleting elements *next is set to NULL by the user and - * av_tree_node_size() will set it to the AVTreeNode which was + * av_tree_insert() will set it to the AVTreeNode which was * used for the removed element. * This allows the use of flat arrays, which have * lower overhead compared to many malloced elements. @@ -95,13 +104,15 @@ void *av_tree_find(const struct AVTreeNode *root, void *key, * return av_tree_insert(rootp, key, cmp, next); * } * @endcode + * @param cmp compare function used to compare elements in the tree, API identical + * to that of Standard C's qsort * @return If no insertion happened, the found element; if an insertion or * removal happened, then either key or NULL will be returned. * Which one it is depends on the tree state and the implementation. You * should make no assumptions that it's one or the other in the code. */ void *av_tree_insert(struct AVTreeNode **rootp, void *key, - int (*cmp)(void *key, const void *b), + int (*cmp)(const void *key, const void *b), struct AVTreeNode **next); void av_tree_destroy(struct AVTreeNode *t); @@ -109,8 +120,8 @@ void av_tree_destroy(struct AVTreeNode *t); /** * Apply enu(opaque, &elem) to all the elements in the tree in a given range. * - * @param cmp a comparison function that returns < 0 for a element below the - * range, > 0 for a element above the range and == 0 for a + * @param cmp a comparison function that returns < 0 for an element below the + * range, > 0 for an element above the range and == 0 for an * element inside the range * * @note The cmp function should use the same ordering used to construct the diff --git a/libavutil/twofish-test.c b/libavutil/twofish-test.c new file mode 100644 index 0000000000..1f22132d6a --- /dev/null +++ b/libavutil/twofish-test.c @@ -0,0 +1,94 @@ +/* + * An implementation of the TwoFish algorithm + * Copyright (c) 2015 Supraja Meedinti + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "twofish.c" +#include "log.h" + +#include <stdio.h> +#include <stdlib.h> + +int main(int argc, char *argv[]) +{ + uint8_t Key[32] = {0x01, 0x23, 0x45, 0x67, 0x89, 0xab, 0xcd, 0xef, 0xfe, 0xdc, 0xba, 0x98, 0x76, 0x54, 0x32, 0x10, 0x00, 0x11, 0x22, 0x33, 0x44, 0x55, 0x66, 0x77, 0x88, 0x99, 0xaa, 0xbb, 0xcc, 0xdd, 0xee, 0xff + }; + const uint8_t rct[6][16] = { + {0x9f, 0x58, 0x9f, 0x5c, 0xf6, 0x12, 0x2c, 0x32, 0xb6, 0xbf, 0xec, 0x2f, 0x2a, 0xe8, 0xc3, 0x5a}, + {0xcf, 0xd1, 0xd2, 0xe5, 0xa9, 0xbe, 0x9c, 0xdf, 0x50, 0x1f, 0x13, 0xb8, 0x92, 0xbd, 0x22, 0x48}, + {0x37, 0x52, 0x7b, 0xe0, 0x05, 0x23, 0x34, 0xb8, 0x9f, 0x0c, 0xfc, 0xca, 0xe8, 0x7c, 0xfa, 0x20}, + {0x5d, 0x9d, 0x4e, 0xef, 0xfa, 0x91, 0x51, 0x57, 0x55, 0x24, 0xf1, 0x15, 0x81, 0x5a, 0x12, 0xe0}, + {0xe7, 0x54, 0x49, 0x21, 0x2b, 0xee, 0xf9, 0xf4, 0xa3, 0x90, 0xbd, 0x86, 0x0a, 0x64, 0x09, 0x41}, + {0x37, 0xfe, 0x26, 0xff, 0x1c, 0xf6, 0x61, 0x75, 0xf5, 0xdd, 0xf4, 0xc3, 0x3b, 0x97, 0xa2, 0x05} + }; + uint8_t temp[32], iv[16], rpt[32] = {0}; + const int kbits[3] = {128, 192, 256}; + int i, j, err = 0; + AVTWOFISH *cs; + cs = av_twofish_alloc(); + if (!cs) + return 1; + for (j = 1; j < 3; j++) { + av_twofish_init(cs, Key, kbits[j]); + av_twofish_crypt(cs, temp, rpt, 1, NULL, 0); + for (i = 0; i < 16; i++) { + if (rct[j][i] != temp[i]) { + av_log(NULL, AV_LOG_ERROR, "%d %02x %02x\n", i, rct[j][i], temp[i]); + err = 1; + } + } + av_twofish_crypt(cs, temp, rct[j], 1, NULL, 1); + for (i = 0; i < 16; i++) { + if (rpt[i] != temp[i]) { + av_log(NULL, AV_LOG_ERROR, "%d %02x %02x\n", i, rpt[i], temp[i]); + err = 1; + } + } + } + for (j = 0; j < 3; j++) { + memset(Key, 0, sizeof(Key)); + memset(rpt, 0, sizeof(rpt)); + for (i = 1; i < 50; i++) { + av_twofish_init(cs, Key, kbits[j]); + av_twofish_crypt(cs, temp, rpt, 1, NULL, 0); + memcpy(Key+16,Key,(kbits[j]-128) >> 3); + memcpy(Key,rpt,16); + memcpy(rpt,temp,16); + } + for (i = 0; i < 16; i++) { + if (rct[3 + j][i] != temp[i]) { + av_log(NULL, AV_LOG_ERROR, "%d %02x %02x\n", i, rct[3 + j][i], temp[i]); + err = 1; + } + } + } + memset(rpt, 0, sizeof(rpt)); + memcpy(iv, "HALLO123HALLO123", 16); + av_twofish_crypt(cs, temp, rpt, 2, iv, 0); + memcpy(iv, "HALLO123HALLO123", 16); + av_twofish_crypt(cs, temp, temp, 2, iv, 1); + for (i = 0; i < 32; i++) { + if (rpt[i] != temp[i]) { + av_log(NULL, AV_LOG_ERROR, "%d %02x %02x\n", i, rpt[i], temp[i]); + err = 1; + } + } + av_free(cs); + return err; +} diff --git a/libavutil/twofish.c b/libavutil/twofish.c new file mode 100644 index 0000000000..d84fa4f363 --- /dev/null +++ b/libavutil/twofish.c @@ -0,0 +1,331 @@ +/* + * An implementation of the TwoFish algorithm + * Copyright (c) 2015 Supraja Meedinti + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ +#include "twofish.h" +#include "common.h" +#include "intreadwrite.h" +#include "attributes.h" + +#define LR(x, n) ((x) << (n) | (x) >> (32 - (n))) +#define RR(x, n) ((x) >> (n) | (x) << (32 - (n))) + +typedef struct AVTWOFISH { + uint32_t K[40]; + uint32_t S[4]; + int ksize; + uint32_t MDS1[256]; + uint32_t MDS2[256]; + uint32_t MDS3[256]; + uint32_t MDS4[256]; +} AVTWOFISH; + +static const uint8_t MD1[256] = { + 0x00, 0x5b, 0xb6, 0xed, 0x05, 0x5e, 0xb3, 0xe8, 0x0a, 0x51, 0xbc, 0xe7, 0x0f, 0x54, 0xb9, 0xe2, + 0x14, 0x4f, 0xa2, 0xf9, 0x11, 0x4a, 0xa7, 0xfc, 0x1e, 0x45, 0xa8, 0xf3, 0x1b, 0x40, 0xad, 0xf6, + 0x28, 0x73, 0x9e, 0xc5, 0x2d, 0x76, 0x9b, 0xc0, 0x22, 0x79, 0x94, 0xcf, 0x27, 0x7c, 0x91, 0xca, + 0x3c, 0x67, 0x8a, 0xd1, 0x39, 0x62, 0x8f, 0xd4, 0x36, 0x6d, 0x80, 0xdb, 0x33, 0x68, 0x85, 0xde, + 0x50, 0x0b, 0xe6, 0xbd, 0x55, 0x0e, 0xe3, 0xb8, 0x5a, 0x01, 0xec, 0xb7, 0x5f, 0x04, 0xe9, 0xb2, + 0x44, 0x1f, 0xf2, 0xa9, 0x41, 0x1a, 0xf7, 0xac, 0x4e, 0x15, 0xf8, 0xa3, 0x4b, 0x10, 0xfd, 0xa6, + 0x78, 0x23, 0xce, 0x95, 0x7d, 0x26, 0xcb, 0x90, 0x72, 0x29, 0xc4, 0x9f, 0x77, 0x2c, 0xc1, 0x9a, + 0x6c, 0x37, 0xda, 0x81, 0x69, 0x32, 0xdf, 0x84, 0x66, 0x3d, 0xd0, 0x8b, 0x63, 0x38, 0xd5, 0x8e, + 0xa0, 0xfb, 0x16, 0x4d, 0xa5, 0xfe, 0x13, 0x48, 0xaa, 0xf1, 0x1c, 0x47, 0xaf, 0xf4, 0x19, 0x42, + 0xb4, 0xef, 0x02, 0x59, 0xb1, 0xea, 0x07, 0x5c, 0xbe, 0xe5, 0x08, 0x53, 0xbb, 0xe0, 0x0d, 0x56, + 0x88, 0xd3, 0x3e, 0x65, 0x8d, 0xd6, 0x3b, 0x60, 0x82, 0xd9, 0x34, 0x6f, 0x87, 0xdc, 0x31, 0x6a, + 0x9c, 0xc7, 0x2a, 0x71, 0x99, 0xc2, 0x2f, 0x74, 0x96, 0xcd, 0x20, 0x7b, 0x93, 0xc8, 0x25, 0x7e, + 0xf0, 0xab, 0x46, 0x1d, 0xf5, 0xae, 0x43, 0x18, 0xfa, 0xa1, 0x4c, 0x17, 0xff, 0xa4, 0x49, 0x12, + 0xe4, 0xbf, 0x52, 0x09, 0xe1, 0xba, 0x57, 0x0c, 0xee, 0xb5, 0x58, 0x03, 0xeb, 0xb0, 0x5d, 0x06, + 0xd8, 0x83, 0x6e, 0x35, 0xdd, 0x86, 0x6b, 0x30, 0xd2, 0x89, 0x64, 0x3f, 0xd7, 0x8c, 0x61, 0x3a, + 0xcc, 0x97, 0x7a, 0x21, 0xc9, 0x92, 0x7f, 0x24, 0xc6, 0x9d, 0x70, 0x2b, 0xc3, 0x98, 0x75, 0x2e +}; + +static const uint8_t MD2[256] = { + 0x00, 0xef, 0xb7, 0x58, 0x07, 0xe8, 0xb0, 0x5f, 0x0e, 0xe1, 0xb9, 0x56, 0x09, 0xe6, 0xbe, 0x51, + 0x1c, 0xf3, 0xab, 0x44, 0x1b, 0xf4, 0xac, 0x43, 0x12, 0xfd, 0xa5, 0x4a, 0x15, 0xfa, 0xa2, 0x4d, + 0x38, 0xd7, 0x8f, 0x60, 0x3f, 0xd0, 0x88, 0x67, 0x36, 0xd9, 0x81, 0x6e, 0x31, 0xde, 0x86, 0x69, + 0x24, 0xcb, 0x93, 0x7c, 0x23, 0xcc, 0x94, 0x7b, 0x2a, 0xc5, 0x9d, 0x72, 0x2d, 0xc2, 0x9a, 0x75, + 0x70, 0x9f, 0xc7, 0x28, 0x77, 0x98, 0xc0, 0x2f, 0x7e, 0x91, 0xc9, 0x26, 0x79, 0x96, 0xce, 0x21, + 0x6c, 0x83, 0xdb, 0x34, 0x6b, 0x84, 0xdc, 0x33, 0x62, 0x8d, 0xd5, 0x3a, 0x65, 0x8a, 0xd2, 0x3d, + 0x48, 0xa7, 0xff, 0x10, 0x4f, 0xa0, 0xf8, 0x17, 0x46, 0xa9, 0xf1, 0x1e, 0x41, 0xae, 0xf6, 0x19, + 0x54, 0xbb, 0xe3, 0x0c, 0x53, 0xbc, 0xe4, 0x0b, 0x5a, 0xb5, 0xed, 0x02, 0x5d, 0xb2, 0xea, 0x05, + 0xe0, 0x0f, 0x57, 0xb8, 0xe7, 0x08, 0x50, 0xbf, 0xee, 0x01, 0x59, 0xb6, 0xe9, 0x06, 0x5e, 0xb1, + 0xfc, 0x13, 0x4b, 0xa4, 0xfb, 0x14, 0x4c, 0xa3, 0xf2, 0x1d, 0x45, 0xaa, 0xf5, 0x1a, 0x42, 0xad, + 0xd8, 0x37, 0x6f, 0x80, 0xdf, 0x30, 0x68, 0x87, 0xd6, 0x39, 0x61, 0x8e, 0xd1, 0x3e, 0x66, 0x89, + 0xc4, 0x2b, 0x73, 0x9c, 0xc3, 0x2c, 0x74, 0x9b, 0xca, 0x25, 0x7d, 0x92, 0xcd, 0x22, 0x7a, 0x95, + 0x90, 0x7f, 0x27, 0xc8, 0x97, 0x78, 0x20, 0xcf, 0x9e, 0x71, 0x29, 0xc6, 0x99, 0x76, 0x2e, 0xc1, + 0x8c, 0x63, 0x3b, 0xd4, 0x8b, 0x64, 0x3c, 0xd3, 0x82, 0x6d, 0x35, 0xda, 0x85, 0x6a, 0x32, 0xdd, + 0xa8, 0x47, 0x1f, 0xf0, 0xaf, 0x40, 0x18, 0xf7, 0xa6, 0x49, 0x11, 0xfe, 0xa1, 0x4e, 0x16, 0xf9, + 0xb4, 0x5b, 0x03, 0xec, 0xb3, 0x5c, 0x04, 0xeb, 0xba, 0x55, 0x0d, 0xe2, 0xbd, 0x52, 0x0a, 0xe5 +}; + +static const uint8_t q0[256] = { + 0xa9, 0x67, 0xb3, 0xe8, 0x04, 0xfd, 0xa3, 0x76, 0x9a, 0x92, 0x80, 0x78, 0xe4, 0xdd, 0xd1, 0x38, + 0x0d, 0xc6, 0x35, 0x98, 0x18, 0xf7, 0xec, 0x6c, 0x43, 0x75, 0x37, 0x26, 0xfa, 0x13, 0x94, 0x48, + 0xf2, 0xd0, 0x8b, 0x30, 0x84, 0x54, 0xdf, 0x23, 0x19, 0x5b, 0x3d, 0x59, 0xf3, 0xae, 0xa2, 0x82, + 0x63, 0x01, 0x83, 0x2e, 0xd9, 0x51, 0x9b, 0x7c, 0xa6, 0xeb, 0xa5, 0xbe, 0x16, 0x0c, 0xe3, 0x61, + 0xc0, 0x8c, 0x3a, 0xf5, 0x73, 0x2c, 0x25, 0x0b, 0xbb, 0x4e, 0x89, 0x6b, 0x53, 0x6a, 0xb4, 0xf1, + 0xe1, 0xe6, 0xbd, 0x45, 0xe2, 0xf4, 0xb6, 0x66, 0xcc, 0x95, 0x03, 0x56, 0xd4, 0x1c, 0x1e, 0xd7, + 0xfb, 0xc3, 0x8e, 0xb5, 0xe9, 0xcf, 0xbf, 0xba, 0xea, 0x77, 0x39, 0xaf, 0x33, 0xc9, 0x62, 0x71, + 0x81, 0x79, 0x09, 0xad, 0x24, 0xcd, 0xf9, 0xd8, 0xe5, 0xc5, 0xb9, 0x4d, 0x44, 0x08, 0x86, 0xe7, + 0xa1, 0x1d, 0xaa, 0xed, 0x06, 0x70, 0xb2, 0xd2, 0x41, 0x7b, 0xa0, 0x11, 0x31, 0xc2, 0x27, 0x90, + 0x20, 0xf6, 0x60, 0xff, 0x96, 0x5c, 0xb1, 0xab, 0x9e, 0x9c, 0x52, 0x1b, 0x5f, 0x93, 0x0a, 0xef, + 0x91, 0x85, 0x49, 0xee, 0x2d, 0x4f, 0x8f, 0x3b, 0x47, 0x87, 0x6d, 0x46, 0xd6, 0x3e, 0x69, 0x64, + 0x2a, 0xce, 0xcb, 0x2f, 0xfc, 0x97, 0x05, 0x7a, 0xac, 0x7f, 0xd5, 0x1a, 0x4b, 0x0e, 0xa7, 0x5a, + 0x28, 0x14, 0x3f, 0x29, 0x88, 0x3c, 0x4c, 0x02, 0xb8, 0xda, 0xb0, 0x17, 0x55, 0x1f, 0x8a, 0x7d, + 0x57, 0xc7, 0x8d, 0x74, 0xb7, 0xc4, 0x9f, 0x72, 0x7e, 0x15, 0x22, 0x12, 0x58, 0x07, 0x99, 0x34, + 0x6e, 0x50, 0xde, 0x68, 0x65, 0xbc, 0xdb, 0xf8, 0xc8, 0xa8, 0x2b, 0x40, 0xdc, 0xfe, 0x32, 0xa4, + 0xca, 0x10, 0x21, 0xf0, 0xd3, 0x5d, 0x0f, 0x00, 0x6f, 0x9d, 0x36, 0x42, 0x4a, 0x5e, 0xc1, 0xe0 +}; + +static const uint8_t q1[256] = { + 0x75, 0xf3, 0xc6, 0xf4, 0xdb, 0x7b, 0xfb, 0xc8, 0x4a, 0xd3, 0xe6, 0x6b, 0x45, 0x7d, 0xe8, 0x4b, + 0xd6, 0x32, 0xd8, 0xfd, 0x37, 0x71, 0xf1, 0xe1, 0x30, 0x0f, 0xf8, 0x1b, 0x87, 0xfa, 0x06, 0x3f, + 0x5e, 0xba, 0xae, 0x5b, 0x8a, 0x00, 0xbc, 0x9d, 0x6d, 0xc1, 0xb1, 0x0e, 0x80, 0x5d, 0xd2, 0xd5, + 0xa0, 0x84, 0x07, 0x14, 0xb5, 0x90, 0x2c, 0xa3, 0xb2, 0x73, 0x4c, 0x54, 0x92, 0x74, 0x36, 0x51, + 0x38, 0xb0, 0xbd, 0x5a, 0xfc, 0x60, 0x62, 0x96, 0x6c, 0x42, 0xf7, 0x10, 0x7c, 0x28, 0x27, 0x8c, + 0x13, 0x95, 0x9c, 0xc7, 0x24, 0x46, 0x3b, 0x70, 0xca, 0xe3, 0x85, 0xcb, 0x11, 0xd0, 0x93, 0xb8, + 0xa6, 0x83, 0x20, 0xff, 0x9f, 0x77, 0xc3, 0xcc, 0x03, 0x6f, 0x08, 0xbf, 0x40, 0xe7, 0x2b, 0xe2, + 0x79, 0x0c, 0xaa, 0x82, 0x41, 0x3a, 0xea, 0xb9, 0xe4, 0x9a, 0xa4, 0x97, 0x7e, 0xda, 0x7a, 0x17, + 0x66, 0x94, 0xa1, 0x1d, 0x3d, 0xf0, 0xde, 0xb3, 0x0b, 0x72, 0xa7, 0x1c, 0xef, 0xd1, 0x53, 0x3e, + 0x8f, 0x33, 0x26, 0x5f, 0xec, 0x76, 0x2a, 0x49, 0x81, 0x88, 0xee, 0x21, 0xc4, 0x1a, 0xeb, 0xd9, + 0xc5, 0x39, 0x99, 0xcd, 0xad, 0x31, 0x8b, 0x01, 0x18, 0x23, 0xdd, 0x1f, 0x4e, 0x2d, 0xf9, 0x48, + 0x4f, 0xf2, 0x65, 0x8e, 0x78, 0x5c, 0x58, 0x19, 0x8d, 0xe5, 0x98, 0x57, 0x67, 0x7f, 0x05, 0x64, + 0xaf, 0x63, 0xb6, 0xfe, 0xf5, 0xb7, 0x3c, 0xa5, 0xce, 0xe9, 0x68, 0x44, 0xe0, 0x4d, 0x43, 0x69, + 0x29, 0x2e, 0xac, 0x15, 0x59, 0xa8, 0x0a, 0x9e, 0x6e, 0x47, 0xdf, 0x34, 0x35, 0x6a, 0xcf, 0xdc, + 0x22, 0xc9, 0xc0, 0x9b, 0x89, 0xd4, 0xed, 0xab, 0x12, 0xa2, 0x0d, 0x52, 0xbb, 0x02, 0x2f, 0xa9, + 0xd7, 0x61, 0x1e, 0xb4, 0x50, 0x04, 0xf6, 0xc2, 0x16, 0x25, 0x86, 0x56, 0x55, 0x09, 0xbe, 0x91 +}; + +struct AVTWOFISH *av_twofish_alloc(void) +{ + return av_mallocz(sizeof(struct AVTWOFISH)); +} + +const int av_twofish_size = sizeof(AVTWOFISH); + +static uint8_t gfmul(uint8_t a, uint8_t b) +{ + uint8_t r = 0, t; + while (a && b) { + if (a & 1) + r = r ^ b; + t = b & 0x80; + b = b << 1; + if (t) + b = b ^ 0x4d; + a = a >> 1; + } + return r; +} + +static uint32_t tf_RS(uint32_t k0, uint32_t k1) +{ + uint8_t s[4], m[8]; + AV_WL32(m, k0); + AV_WL32(m + 4, k1); + s[0] = gfmul(0x01, m[0]) ^ gfmul(0xa4, m[1]) ^ gfmul(0x55, m[2]) ^ gfmul(0x87, m[3]) ^ gfmul(0x5a, m[4]) ^ gfmul(0x58, m[5]) ^ gfmul(0xdb, m[6]) ^ gfmul(0x9e, m[7]); + s[1] = gfmul(0xa4, m[0]) ^ gfmul(0x56, m[1]) ^ gfmul(0x82, m[2]) ^ gfmul(0xf3, m[3]) ^ gfmul(0x1e, m[4]) ^ gfmul(0xc6, m[5]) ^ gfmul(0x68, m[6]) ^ gfmul(0xe5, m[7]); + s[2] = gfmul(0x02, m[0]) ^ gfmul(0xa1, m[1]) ^ gfmul(0xfc, m[2]) ^ gfmul(0xc1, m[3]) ^ gfmul(0x47, m[4]) ^ gfmul(0xae, m[5]) ^ gfmul(0x3d, m[6]) ^ gfmul(0x19, m[7]); + s[3] = gfmul(0xa4, m[0]) ^ gfmul(0x55, m[1]) ^ gfmul(0x87, m[2]) ^ gfmul(0x5a, m[3]) ^ gfmul(0x58, m[4]) ^ gfmul(0xdb, m[5]) ^ gfmul(0x9e, m[6]) ^ gfmul(0x03, m[7]); + return AV_RL32(s); +} + +static void tf_h0(uint8_t y[4], uint32_t L[4], int k) +{ + uint8_t l[4]; + if (k == 4) { + AV_WL32(l, L[3]); + y[0] = q1[y[0]] ^ l[0]; + y[1] = q0[y[1]] ^ l[1]; + y[2] = q0[y[2]] ^ l[2]; + y[3] = q1[y[3]] ^ l[3]; + } + if (k >= 3) { + AV_WL32(l, L[2]); + y[0] = q1[y[0]] ^ l[0]; + y[1] = q1[y[1]] ^ l[1]; + y[2] = q0[y[2]] ^ l[2]; + y[3] = q0[y[3]] ^ l[3]; + } + AV_WL32(l, L[1]); + y[0] = q1[q0[q0[y[0]] ^ l[0]] ^ (L[0] & 0xff)]; + y[1] = q0[q0[q1[y[1]] ^ l[1]] ^ ((L[0] >> 8) & 0xff)]; + y[2] = q1[q1[q0[y[2]] ^ l[2]] ^ ((L[0] >> 16) & 0xff)]; + y[3] = q0[q1[q1[y[3]] ^ l[3]] ^ (L[0] >> 24)]; +} + +static uint32_t tf_h(uint32_t X, uint32_t L[4], int k) +{ + uint8_t y[4], l[4]; + AV_WL32(y, X); + tf_h0(y, L, k); + + l[0] = y[0] ^ MD2[y[1]] ^ MD1[y[2]] ^ MD1[y[3]]; + l[1] = MD1[y[0]] ^ MD2[y[1]] ^ MD2[y[2]] ^ y[3]; + l[2] = MD2[y[0]] ^ MD1[y[1]] ^ y[2] ^ MD2[y[3]]; + l[3] = MD2[y[0]] ^ y[1] ^ MD2[y[2]] ^ MD1[y[3]]; + + return AV_RL32(l); +} + +static uint32_t MDS_mul(AVTWOFISH *cs, uint32_t X) +{ + return cs->MDS1[(X) & 0xff] ^ cs->MDS2[((X) >> 8) & 0xff] ^ cs->MDS3[((X) >> 16) & 0xff] ^ cs->MDS4[(X) >> 24]; +} + +static void precomputeMDS(AVTWOFISH *cs) +{ + uint8_t y[4]; + int i; + for (i = 0; i < 256; i++) { + y[0] = y[1] = y[2] = y[3] = i; + tf_h0(y, cs->S, cs->ksize); + cs->MDS1[i] = ((uint32_t)y[0]) ^ ((uint32_t)MD1[y[0]] << 8) ^ ((uint32_t)MD2[y[0]] << 16) ^ ((uint32_t)MD2[y[0]] << 24); + cs->MDS2[i] = ((uint32_t)MD2[y[1]]) ^ ((uint32_t)MD2[y[1]] << 8) ^ ((uint32_t)MD1[y[1]] << 16) ^ ((uint32_t)y[1] << 24); + cs->MDS3[i] = ((uint32_t)MD1[y[2]]) ^ ((uint32_t)MD2[y[2]] << 8) ^ ((uint32_t)y[2] << 16) ^ ((uint32_t)MD2[y[2]] << 24); + cs->MDS4[i] = ((uint32_t)MD1[y[3]]) ^ ((uint32_t)y[3] << 8) ^ ((uint32_t)MD2[y[3]] << 16) ^ ((uint32_t)MD1[y[3]] << 24); + } +} + +static void twofish_encrypt(AVTWOFISH *cs, uint8_t *dst, const uint8_t *src) +{ + uint32_t P[4], t0, t1; + int i; + P[0] = AV_RL32(src) ^ cs->K[0]; + P[1] = AV_RL32(src + 4) ^ cs->K[1]; + P[2] = AV_RL32(src + 8) ^ cs->K[2]; + P[3] = AV_RL32(src + 12) ^ cs->K[3]; + for (i = 0; i < 16; i += 2) { + t0 = MDS_mul(cs, P[0]); + t1 = MDS_mul(cs, LR(P[1], 8)); + P[2] = RR(P[2] ^ (t0 + t1 + cs->K[2 * i + 8]), 1); + P[3] = LR(P[3], 1) ^ (t0 + 2 * t1 + cs->K[2 * i + 9]); + t0 = MDS_mul(cs, P[2]); + t1 = MDS_mul(cs, LR(P[3], 8)); + P[0] = RR(P[0] ^ (t0 + t1 + cs->K[2 * i + 10]), 1); + P[1] = LR(P[1], 1) ^ (t0 + 2 * t1 + cs->K[2 * i + 11]); + } + P[2] ^= cs->K[4]; + P[3] ^= cs->K[5]; + P[0] ^= cs->K[6]; + P[1] ^= cs->K[7]; + AV_WL32(dst, P[2]); + AV_WL32(dst + 4, P[3]); + AV_WL32(dst + 8, P[0]); + AV_WL32(dst + 12, P[1]); +} + +static void twofish_decrypt(AVTWOFISH *cs, uint8_t *dst, const uint8_t *src, uint8_t *iv) +{ + uint32_t P[4], t0, t1; + int i; + P[2] = AV_RL32(src) ^ cs->K[4]; + P[3] = AV_RL32(src + 4) ^ cs->K[5]; + P[0] = AV_RL32(src + 8) ^ cs->K[6]; + P[1] = AV_RL32(src + 12) ^ cs->K[7]; + for (i = 15; i >= 0; i -= 2) { + t0 = MDS_mul(cs, P[2]); + t1 = MDS_mul(cs, LR(P[3], 8)); + P[0] = LR(P[0], 1) ^ (t0 + t1 + cs->K[2 * i + 8]); + P[1] = RR(P[1] ^ (t0 + 2 * t1 + cs->K[2 * i + 9]), 1); + t0 = MDS_mul(cs, P[0]); + t1 = MDS_mul(cs, LR(P[1], 8)); + P[2] = LR(P[2], 1) ^ (t0 + t1 + cs->K[2 * i + 6]); + P[3] = RR(P[3] ^ (t0 + 2 * t1 + cs->K[2 * i + 7]), 1); + } + P[0] ^= cs->K[0]; + P[1] ^= cs->K[1]; + P[2] ^= cs->K[2]; + P[3] ^= cs->K[3]; + if (iv) { + P[0] ^= AV_RL32(iv); + P[1] ^= AV_RL32(iv + 4); + P[2] ^= AV_RL32(iv + 8); + P[3] ^= AV_RL32(iv + 12); + memcpy(iv, src, 16); + } + AV_WL32(dst, P[2]); + AV_WL32(dst + 4, P[3]); + AV_WL32(dst + 8, P[0]); + AV_WL32(dst + 12, P[1]); +} + +av_cold int av_twofish_init(AVTWOFISH *cs, const uint8_t *key, int key_bits) +{ + int i; + uint8_t keypad[32]; + uint32_t Key[8], Me[4], Mo[4], A, B; + const uint32_t rho = 0x01010101; + if (key_bits < 0) + return AVERROR(EINVAL); + if (key_bits <= 128) { + cs->ksize = 2; + } else if (key_bits <= 192) { + cs->ksize = 3; + } else { + cs->ksize = 4; + } + memset(keypad, 0, sizeof(keypad)); + if (key_bits <= 256) { + memcpy(keypad, key, key_bits >> 3); + } else { + memcpy(keypad, key, 32); + } + for (i = 0; i < 2 * cs->ksize ; i++) + Key[i] = AV_RL32(keypad + 4 * i); + for (i = 0; i < cs->ksize; i++) { + Me[i] = Key[2 * i]; + Mo[i] = Key[2 * i + 1]; + cs->S[cs->ksize - i - 1] = tf_RS(Me[i], Mo[i]); + } + precomputeMDS(cs); + for (i = 0; i < 20; i++) { + A = tf_h((2 * i) * rho, Me, cs->ksize); + B = tf_h((2 * i + 1) * rho, Mo, cs->ksize); + B = LR(B, 8); + cs->K[2 * i] = A + B; + cs->K[2 * i + 1] = LR((A + (2 * B)), 9); + } + if (cs->ksize << 6 != key_bits) { + return 1; + } else { + return 0; + } +} + +void av_twofish_crypt(AVTWOFISH *cs, uint8_t *dst, const uint8_t *src, int count, uint8_t *iv, int decrypt) +{ + int i; + while (count--) { + if (decrypt) { + twofish_decrypt(cs, dst, src, iv); + } else { + if (iv) { + for (i = 0; i < 16; i++) + dst[i] = src[i] ^ iv[i]; + twofish_encrypt(cs, dst, dst); + memcpy(iv, dst, 16); + } else { + twofish_encrypt(cs, dst, src); + } + } + src = src + 16; + dst = dst + 16; + } +} diff --git a/libavutil/twofish.h b/libavutil/twofish.h new file mode 100644 index 0000000000..813cfecdf8 --- /dev/null +++ b/libavutil/twofish.h @@ -0,0 +1,70 @@ +/* + * An implementation of the TwoFish algorithm + * Copyright (c) 2015 Supraja Meedinti + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#ifndef AVUTIL_TWOFISH_H +#define AVUTIL_TWOFISH_H + +#include <stdint.h> + + +/** + * @file + * @brief Public header for libavutil TWOFISH algorithm + * @defgroup lavu_twofish TWOFISH + * @ingroup lavu_crypto + * @{ + */ + +extern const int av_twofish_size; + +struct AVTWOFISH; + +/** + * Allocate an AVTWOFISH context + * To free the struct: av_free(ptr) + */ +struct AVTWOFISH *av_twofish_alloc(void); + +/** + * Initialize an AVTWOFISH context. + * + * @param ctx an AVTWOFISH context + * @param key a key of size ranging from 1 to 32 bytes used for encryption/decryption + * @param key_bits number of keybits: 128, 192, 256 If less than the required, padded with zeroes to nearest valid value; return value is 0 if key_bits is 128/192/256, -1 if less than 0, 1 otherwise + */ +int av_twofish_init(struct AVTWOFISH *ctx, const uint8_t *key, int key_bits); + +/** + * Encrypt or decrypt a buffer using a previously initialized context + * + * @param ctx an AVTWOFISH context + * @param dst destination array, can be equal to src + * @param src source array, can be equal to dst + * @param count number of 16 byte blocks + * @paran iv initialization vector for CBC mode, NULL for ECB mode + * @param decrypt 0 for encryption, 1 for decryption + */ +void av_twofish_crypt(struct AVTWOFISH *ctx, uint8_t *dst, const uint8_t *src, int count, uint8_t* iv, int decrypt); + +/** + * @} + */ +#endif /* AVUTIL_TWOFISH_H */ diff --git a/libavutil/utf8-test.c b/libavutil/utf8-test.c new file mode 100644 index 0000000000..37a2802b5f --- /dev/null +++ b/libavutil/utf8-test.c @@ -0,0 +1,71 @@ +/* + * Copyright (c) 2013 Stefano Sabatini + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include <stdio.h> + +#include "libavutil/avstring.h" +#include "libavutil/file.h" + +static void print_sequence(const char *p, int l, int indent) +{ + int i; + for (i = 0; i < l; i++) + printf("%02X", (uint8_t)p[i]); + printf("%*s", indent-l*2, ""); +} + +int main(int argc, char **argv) +{ + int ret; + char *filename = argv[1]; + uint8_t *file_buf; + size_t file_buf_size; + uint32_t code; + const uint8_t *p, *endp; + + ret = av_file_map(filename, &file_buf, &file_buf_size, 0, NULL); + if (ret < 0) + return 1; + + p = file_buf; + endp = file_buf + file_buf_size; + while (p < endp) { + int l, r; + const uint8_t *p0 = p; + code = UINT32_MAX; + r = av_utf8_decode(&code, &p, endp, 0); + l = (int)(p-p0); + print_sequence(p0, l, 20); + if (code != UINT32_MAX) { + printf("%-10d 0x%-10X %-5d ", code, code, l); + if (r >= 0) { + if (*p0 == '\n') printf("\\n\n"); + else printf ("%.*s\n", l, p0); + } else { + printf("invalid code range\n"); + } + } else { + printf("invalid sequence\n"); + } + } + + av_file_unmap(file_buf, file_buf_size); + return 0; +} diff --git a/libavutil/utils.c b/libavutil/utils.c index 8dcf3b6774..f409f32216 100644 --- a/libavutil/utils.c +++ b/libavutil/utils.c @@ -1,50 +1,91 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ #include "config.h" #include "avutil.h" - -#include "avversion.h" +#include "avassert.h" +#include "samplefmt.h" +#include "internal.h" /** * @file * various utility functions */ +#include "libavutil/ffversion.h" +const char av_util_ffversion[] = "FFmpeg version " FFMPEG_VERSION; + const char *av_version_info(void) { - return LIBAV_VERSION; + return FFMPEG_VERSION; } unsigned avutil_version(void) { + static int checks_done; + if (checks_done) + return LIBAVUTIL_VERSION_INT; + +#if FF_API_VDPAU + av_assert0(AV_PIX_FMT_VDA_VLD == 81); //check if the pix fmt enum has not had anything inserted or removed by mistake +#endif + av_assert0(AV_SAMPLE_FMT_DBLP == 9); + av_assert0(AVMEDIA_TYPE_ATTACHMENT == 4); + av_assert0(AV_PICTURE_TYPE_BI == 7); + av_assert0(LIBAVUTIL_VERSION_MICRO >= 100); + av_assert0(HAVE_MMX2 == HAVE_MMXEXT); + + av_assert0(((size_t)-1) > 0); // C guarantees this but if false on a platform we care about revert at least b284e1ffe343d6697fb950d1ee517bafda8a9844 + + if (av_sat_dadd32(1, 2) != 5) { + av_log(NULL, AV_LOG_FATAL, "Libavutil has been built with a broken binutils, please upgrade binutils and rebuild\n"); + abort(); + } + + if (llrint(1LL<<60) != 1LL<<60) { + av_log(NULL, AV_LOG_ERROR, "Libavutil has been linked to a broken llrint()\n"); + } + + checks_done = 1; return LIBAVUTIL_VERSION_INT; } const char *avutil_configuration(void) { - return LIBAV_CONFIGURATION; + return FFMPEG_CONFIGURATION; } const char *avutil_license(void) { #define LICENSE_PREFIX "libavutil license: " - return LICENSE_PREFIX LIBAV_LICENSE + sizeof(LICENSE_PREFIX) - 1; + return LICENSE_PREFIX FFMPEG_LICENSE + sizeof(LICENSE_PREFIX) - 1; +} + +const char *av_get_media_type_string(enum AVMediaType media_type) +{ + switch (media_type) { + case AVMEDIA_TYPE_VIDEO: return "video"; + case AVMEDIA_TYPE_AUDIO: return "audio"; + case AVMEDIA_TYPE_DATA: return "data"; + case AVMEDIA_TYPE_SUBTITLE: return "subtitle"; + case AVMEDIA_TYPE_ATTACHMENT: return "attachment"; + default: return NULL; + } } char av_get_picture_type_char(enum AVPictureType pict_type) @@ -61,6 +102,25 @@ char av_get_picture_type_char(enum AVPictureType pict_type) } } +unsigned av_int_list_length_for_size(unsigned elsize, + const void *list, uint64_t term) +{ + unsigned i; + + if (!list) + return 0; +#define LIST_LENGTH(type) \ + { type t = term, *l = (type *)list; for (i = 0; l[i] != t; i++); } + switch (elsize) { + case 1: LIST_LENGTH(uint8_t); break; + case 2: LIST_LENGTH(uint16_t); break; + case 4: LIST_LENGTH(uint32_t); break; + case 8: LIST_LENGTH(uint64_t); break; + default: av_assert0(!"valid element size"); + } + return i; +} + AVRational av_get_time_base_q(void) { return (AVRational){1, AV_TIME_BASE}; diff --git a/libavutil/version.h b/libavutil/version.h index 186ebd8eb5..35b6374f36 100644 --- a/libavutil/version.h +++ b/libavutil/version.h @@ -1,18 +1,20 @@ /* - * This file is part of Libav. + * copyright (c) 2003 Fabrice Bellard * - * Libav is free software; you can redistribute it and/or + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -22,7 +24,7 @@ #include "macros.h" /** - * @defgroup version_utils Library Version Macros + * @addtogroup version_utils * * Useful to check and match library version in order to maintain * backward compatibility. @@ -30,11 +32,19 @@ * @{ */ -#define AV_VERSION_INT(a, b, c) (a<<16 | b<<8 | c) +#define AV_VERSION_INT(a, b, c) ((a)<<16 | (b)<<8 | (c)) #define AV_VERSION_DOT(a, b, c) a ##.## b ##.## c #define AV_VERSION(a, b, c) AV_VERSION_DOT(a, b, c) /** + * Extract version components from the full ::AV_VERSION_INT int as returned + * by functions like ::avformat_version() and ::avcodec_version() + */ +#define AV_VERSION_MAJOR(a) ((a) >> 16) +#define AV_VERSION_MINOR(a) (((a) & 0x00FF00) >> 8) +#define AV_VERSION_MICRO(a) ((a) & 0xFF) + +/** * @} */ @@ -53,9 +63,9 @@ * @{ */ -#define LIBAVUTIL_VERSION_MAJOR 55 -#define LIBAVUTIL_VERSION_MINOR 10 -#define LIBAVUTIL_VERSION_MICRO 0 +#define LIBAVUTIL_VERSION_MAJOR 55 +#define LIBAVUTIL_VERSION_MINOR 24 +#define LIBAVUTIL_VERSION_MICRO 100 #define LIBAVUTIL_VERSION_INT AV_VERSION_INT(LIBAVUTIL_VERSION_MAJOR, \ LIBAVUTIL_VERSION_MINOR, \ @@ -75,6 +85,10 @@ * dropped at a future version bump. The defines themselves are not part of * the public API and may change, break or disappear at any time. * + * @note, when bumping the major version it is recommended to manually + * disable each FF_API_* in its own commit instead of disabling them all + * at once through the bump. This improves the git bisect-ability of the change. + * * @{ */ @@ -90,8 +104,11 @@ #ifndef FF_API_DLOG #define FF_API_DLOG (LIBAVUTIL_VERSION_MAJOR < 56) #endif -#ifndef FF_API_CRYPTO_CONTEXT -#define FF_API_CRYPTO_CONTEXT (LIBAVUTIL_VERSION_MAJOR < 56) +#ifndef FF_API_VAAPI +#define FF_API_VAAPI (LIBAVUTIL_VERSION_MAJOR < 56) +#endif +#ifndef FF_API_FRAME_QP +#define FF_API_FRAME_QP (LIBAVUTIL_VERSION_MAJOR < 56) #endif #ifndef FF_API_PLUS1_MINUS1 #define FF_API_PLUS1_MINUS1 (LIBAVUTIL_VERSION_MAJOR < 56) @@ -99,8 +116,8 @@ #ifndef FF_API_ERROR_FRAME #define FF_API_ERROR_FRAME (LIBAVUTIL_VERSION_MAJOR < 56) #endif -#ifndef FF_API_VAAPI -#define FF_API_VAAPI (LIBAVUTIL_VERSION_MAJOR < 56) +#ifndef FF_API_CRC_BIG_TABLE +#define FF_API_CRC_BIG_TABLE (LIBAVUTIL_VERSION_MAJOR < 56) #endif diff --git a/libavutil/wchar_filename.h b/libavutil/wchar_filename.h index 2781773dee..2ade321bed 100644 --- a/libavutil/wchar_filename.h +++ b/libavutil/wchar_filename.h @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -23,6 +23,7 @@ #include <windows.h> #include "mem.h" +av_warn_unused_result static inline int utf8towchar(const char *filename_utf8, wchar_t **filename_w) { int num_chars; @@ -31,7 +32,7 @@ static inline int utf8towchar(const char *filename_utf8, wchar_t **filename_w) *filename_w = NULL; return 0; } - *filename_w = av_mallocz(sizeof(wchar_t) * num_chars); + *filename_w = (wchar_t *)av_mallocz_array(num_chars, sizeof(wchar_t)); if (!*filename_w) { errno = ENOMEM; return -1; diff --git a/libavutil/x86/Makefile b/libavutil/x86/Makefile index 1e19082233..94d8832062 100644 --- a/libavutil/x86/Makefile +++ b/libavutil/x86/Makefile @@ -1,8 +1,16 @@ OBJS += x86/cpu.o \ + x86/fixed_dsp_init.o \ x86/float_dsp_init.o \ x86/lls_init.o \ +OBJS-$(CONFIG_PIXELUTILS) += x86/pixelutils_init.o \ + +EMMS_OBJS_$(HAVE_MMX_INLINE)_$(HAVE_MMX_EXTERNAL)_$(HAVE_MM_EMPTY) = x86/emms.o + YASM-OBJS += x86/cpuid.o \ - x86/emms.o \ + $(EMMS_OBJS__yes_) \ + x86/fixed_dsp.o \ x86/float_dsp.o \ x86/lls.o \ + +YASM-OBJS-$(CONFIG_PIXELUTILS) += x86/pixelutils.o \ diff --git a/libavutil/x86/asm.h b/libavutil/x86/asm.h index db5f3d5ac5..109b65e542 100644 --- a/libavutil/x86/asm.h +++ b/libavutil/x86/asm.h @@ -1,20 +1,20 @@ /* * copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -38,7 +38,8 @@ typedef struct ymm_reg { uint64_t a, b, c, d; } ymm_reg; # define PTR_SIZE "8" typedef int64_t x86_reg; -# define REG_SP "rsp" +/* REG_SP is defined in Solaris sys headers, so use REG_sp */ +# define REG_sp "rsp" # define REG_BP "rbp" # define REGBP rbp # define REGa rax @@ -59,7 +60,7 @@ typedef int64_t x86_reg; # define PTR_SIZE "4" typedef int32_t x86_reg; -# define REG_SP "esp" +# define REG_sp "esp" # define REG_BP "ebp" # define REGBP ebp # define REGa eax @@ -108,6 +109,46 @@ typedef int x86_reg; # define LOCAL_MANGLE(a) #a #endif -#define MANGLE(a) EXTERN_PREFIX LOCAL_MANGLE(a) +#if HAVE_INLINE_ASM_DIRECT_SYMBOL_REFS +# define MANGLE(a) EXTERN_PREFIX LOCAL_MANGLE(a) +# define NAMED_CONSTRAINTS_ADD(...) +# define NAMED_CONSTRAINTS(...) +# define NAMED_CONSTRAINTS_ARRAY_ADD(...) +# define NAMED_CONSTRAINTS_ARRAY(...) +#else + /* When direct symbol references are used in code passed to a compiler that does not support them + * then these references need to be converted to named asm constraints instead. + * Instead of returning a direct symbol MANGLE now returns a named constraint for that specific symbol. + * In order for this to work there must also be a corresponding entry in the asm-interface. To add this + * entry use the macro NAMED_CONSTRAINTS() and pass in a list of each symbol reference used in the + * corresponding block of code. (e.g. NAMED_CONSTRAINTS(var1,var2,var3) where var1 is the first symbol etc. ). + * If there are already existing constraints then use NAMED_CONSTRAINTS_ADD to add to the existing constraint list. + */ +# define MANGLE(a) "%["#a"]" + // Intel/MSVC does not correctly expand va-args so we need a rather ugly hack in order to get it to work +# define FE_0(P,X) P(X) +# define FE_1(P,X,X1) P(X), FE_0(P,X1) +# define FE_2(P,X,X1,X2) P(X), FE_1(P,X1,X2) +# define FE_3(P,X,X1,X2,X3) P(X), FE_2(P,X1,X2,X3) +# define FE_4(P,X,X1,X2,X3,X4) P(X), FE_3(P,X1,X2,X3,X4) +# define FE_5(P,X,X1,X2,X3,X4,X5) P(X), FE_4(P,X1,X2,X3,X4,X5) +# define FE_6(P,X,X1,X2,X3,X4,X5,X6) P(X), FE_5(P,X1,X2,X3,X4,X5,X6) +# define FE_7(P,X,X1,X2,X3,X4,X5,X6,X7) P(X), FE_6(P,X1,X2,X3,X4,X5,X6,X7) +# define FE_8(P,X,X1,X2,X3,X4,X5,X6,X7,X8) P(X), FE_7(P,X1,X2,X3,X4,X5,X6,X7,X8) +# define FE_9(P,X,X1,X2,X3,X4,X5,X6,X7,X8,X9) P(X), FE_8(P,X1,X2,X3,X4,X5,X6,X7,X8,X9) +# define GET_FE_IMPL(_0,_1,_2,_3,_4,_5,_6,_7,_8,_9,NAME,...) NAME +# define GET_FE(A) GET_FE_IMPL A +# define GET_FE_GLUE(x, y) x y +# define FOR_EACH_VA(P,...) GET_FE_GLUE(GET_FE((__VA_ARGS__,FE_9,FE_8,FE_7,FE_6,FE_5,FE_4,FE_3,FE_2,FE_1,FE_0)), (P,__VA_ARGS__)) +# define NAME_CONSTRAINT(x) [x] "m"(x) + // Parameters are a list of each symbol reference required +# define NAMED_CONSTRAINTS_ADD(...) , FOR_EACH_VA(NAME_CONSTRAINT,__VA_ARGS__) + // Same but without comma for when there are no previously defined constraints +# define NAMED_CONSTRAINTS(...) FOR_EACH_VA(NAME_CONSTRAINT,__VA_ARGS__) + // Same as above NAMED_CONSTRAINTS except used for passing arrays/pointers instead of normal variables +# define NAME_CONSTRAINT_ARRAY(x) [x] "m"(*x) +# define NAMED_CONSTRAINTS_ARRAY_ADD(...) , FOR_EACH_VA(NAME_CONSTRAINT_ARRAY,__VA_ARGS__) +# define NAMED_CONSTRAINTS_ARRAY(...) FOR_EACH_VA(NAME_CONSTRAINT_ARRAY,__VA_ARGS__) +#endif #endif /* AVUTIL_X86_ASM_H */ diff --git a/libavutil/x86/bswap.h b/libavutil/x86/bswap.h index c73be9af81..ffa59e4c82 100644 --- a/libavutil/x86/bswap.h +++ b/libavutil/x86/bswap.h @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -25,21 +25,47 @@ #define AVUTIL_X86_BSWAP_H #include <stdint.h> +#if defined(_MSC_VER) +#include <intrin.h> +#endif #include "config.h" #include "libavutil/attributes.h" -#if HAVE_INLINE_ASM +#if defined(_MSC_VER) + +#define av_bswap16 av_bswap16 +static av_always_inline av_const uint16_t av_bswap16(uint16_t x) +{ + return _rotr16(x, 8); +} + +#define av_bswap32 av_bswap32 +static av_always_inline av_const uint32_t av_bswap32(uint32_t x) +{ + return _byteswap_ulong(x); +} + +#if ARCH_X86_64 +#define av_bswap64 av_bswap64 +static inline uint64_t av_const av_bswap64(uint64_t x) +{ + return _byteswap_uint64(x); +} +#endif + + +#elif HAVE_INLINE_ASM -#if !AV_GCC_VERSION_AT_LEAST(4,1) +#if AV_GCC_VERSION_AT_MOST(4,0) #define av_bswap16 av_bswap16 static av_always_inline av_const unsigned av_bswap16(unsigned x) { __asm__("rorw $8, %w0" : "+r"(x)); return x; } -#endif /* !AV_GCC_VERSION_AT_LEAST(4,1) */ +#endif /* AV_GCC_VERSION_AT_MOST(4,0) */ -#if !AV_GCC_VERSION_AT_LEAST(4,5) +#if AV_GCC_VERSION_AT_MOST(4,4) || defined(__INTEL_COMPILER) #define av_bswap32 av_bswap32 static av_always_inline av_const uint32_t av_bswap32(uint32_t x) { @@ -55,7 +81,7 @@ static inline uint64_t av_const av_bswap64(uint64_t x) return x; } #endif -#endif /* !AV_GCC_VERSION_AT_LEAST(4,5) */ +#endif /* AV_GCC_VERSION_AT_MOST(4,4) */ #endif /* HAVE_INLINE_ASM */ #endif /* AVUTIL_X86_BSWAP_H */ diff --git a/libavutil/x86/cpu.c b/libavutil/x86/cpu.c index 098ccf7004..bb63daac3d 100644 --- a/libavutil/x86/cpu.c +++ b/libavutil/x86/cpu.c @@ -3,20 +3,20 @@ * (c)1997-99 by H. Dietz and R. Fisher * Converted to C and improved by Fabrice Bellard. * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -45,7 +45,7 @@ "cpuid \n\t" \ "xchg %%"REG_b", %%"REG_S \ : "=a" (eax), "=S" (ebx), "=c" (ecx), "=d" (edx) \ - : "0" (index)) + : "0" (index), "2"(0)) #define xgetbv(index, eax, edx) \ __asm__ (".byte 0x0f, 0x01, 0xd0" : "=a"(eax), "=d"(edx) : "c" (index)) @@ -126,6 +126,8 @@ int ff_get_cpu_flags_x86(void) rval |= AV_CPU_FLAG_SSE4; if (ecx & 0x00100000 ) rval |= AV_CPU_FLAG_SSE42; + if (ecx & 0x01000000 ) + rval |= AV_CPU_FLAG_AESNI; #if HAVE_AVX /* Check OXSAVE and AVX bits */ if ((ecx & 0x18000000) == 0x18000000) { @@ -143,7 +145,7 @@ int ff_get_cpu_flags_x86(void) if (max_std_level >= 7) { cpuid(7, eax, ebx, ecx, edx); #if HAVE_AVX2 - if (ebx & 0x00000020) + if ((rval & AV_CPU_FLAG_AVX) && (ebx & 0x00000020)) rval |= AV_CPU_FLAG_AVX2; #endif /* HAVE_AVX2 */ /* BMI1/2 don't need OS support */ @@ -180,13 +182,11 @@ int ff_get_cpu_flags_x86(void) /* Similar to the above but for AVX functions on AMD processors. This is necessary only for functions using YMM registers on Bulldozer - based CPUs as they lack 256-bits execution units. SSE/AVX functions - using XMM registers are always faster on them. + and Jaguar based CPUs as they lack 256-bits execution units. SSE/AVX + functions using XMM registers are always faster on them. AV_CPU_FLAG_AVX and AV_CPU_FLAG_AVXSLOW are both set so that AVX is - used unless explicitly disabled by checking AV_CPU_FLAG_AVXSLOW. - TODO: Confirm if Excavator is affected or not by this once it's - released, and update the check if necessary. Same for btver2. */ - if (family == 0x15 && (rval & AV_CPU_FLAG_AVX)) + used unless explicitly disabled by checking AV_CPU_FLAG_AVXSLOW. */ + if ((family == 0x15 || family == 0x16) && (rval & AV_CPU_FLAG_AVX)) rval |= AV_CPU_FLAG_AVXSLOW; } diff --git a/libavutil/x86/cpu.h b/libavutil/x86/cpu.h index 0695436548..f171037f1c 100644 --- a/libavutil/x86/cpu.h +++ b/libavutil/x86/cpu.h @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -47,6 +47,7 @@ #define X86_FMA3(flags) CPUEXT(flags, FMA3) #define X86_FMA4(flags) CPUEXT(flags, FMA4) #define X86_AVX2(flags) CPUEXT(flags, AVX2) +#define X86_AESNI(flags) CPUEXT(flags, AESNI) #define EXTERNAL_AMD3DNOW(flags) CPUEXT_SUFFIX(flags, _EXTERNAL, AMD3DNOW) #define EXTERNAL_AMD3DNOWEXT(flags) CPUEXT_SUFFIX(flags, _EXTERNAL, AMD3DNOWEXT) @@ -67,8 +68,13 @@ #define EXTERNAL_AVX_SLOW(flags) CPUEXT_SUFFIX_SLOW(flags, _EXTERNAL, AVX) #define EXTERNAL_XOP(flags) CPUEXT_SUFFIX(flags, _EXTERNAL, XOP) #define EXTERNAL_FMA3(flags) CPUEXT_SUFFIX(flags, _EXTERNAL, FMA3) +#define EXTERNAL_FMA3_FAST(flags) CPUEXT_SUFFIX_FAST2(flags, _EXTERNAL, FMA3, AVX) +#define EXTERNAL_FMA3_SLOW(flags) CPUEXT_SUFFIX_SLOW2(flags, _EXTERNAL, FMA3, AVX) #define EXTERNAL_FMA4(flags) CPUEXT_SUFFIX(flags, _EXTERNAL, FMA4) #define EXTERNAL_AVX2(flags) CPUEXT_SUFFIX(flags, _EXTERNAL, AVX2) +#define EXTERNAL_AVX2_FAST(flags) CPUEXT_SUFFIX_FAST2(flags, _EXTERNAL, AVX2, AVX) +#define EXTERNAL_AVX2_SLOW(flags) CPUEXT_SUFFIX_SLOW2(flags, _EXTERNAL, AVX2, AVX) +#define EXTERNAL_AESNI(flags) CPUEXT_SUFFIX(flags, _EXTERNAL, AESNI) #define INLINE_AMD3DNOW(flags) CPUEXT_SUFFIX(flags, _INLINE, AMD3DNOW) #define INLINE_AMD3DNOWEXT(flags) CPUEXT_SUFFIX(flags, _INLINE, AMD3DNOWEXT) @@ -91,6 +97,7 @@ #define INLINE_FMA3(flags) CPUEXT_SUFFIX(flags, _INLINE, FMA3) #define INLINE_FMA4(flags) CPUEXT_SUFFIX(flags, _INLINE, FMA4) #define INLINE_AVX2(flags) CPUEXT_SUFFIX(flags, _INLINE, AVX2) +#define INLINE_AESNI(flags) CPUEXT_SUFFIX(flags, _INLINE, AESNI) void ff_cpu_cpuid(int index, int *eax, int *ebx, int *ecx, int *edx); void ff_cpu_xgetbv(int op, int *eax, int *edx); diff --git a/libavutil/x86/cpuid.asm b/libavutil/x86/cpuid.asm index 1cb8e94ea3..c3f7866ec7 100644 --- a/libavutil/x86/cpuid.asm +++ b/libavutil/x86/cpuid.asm @@ -4,20 +4,20 @@ ;* Authors: Loren Merritt <lorenm@u.washington.edu> ;* Fiona Glaser <fiona@x264.com> ;* -;* This file is part of Libav. +;* This file is part of FFmpeg. ;* -;* Libav is free software; you can redistribute it and/or +;* FFmpeg is free software; you can redistribute it and/or ;* modify it under the terms of the GNU Lesser General Public ;* License as published by the Free Software Foundation; either ;* version 2.1 of the License, or (at your option) any later version. ;* -;* Libav is distributed in the hope that it will be useful, +;* FFmpeg is distributed in the hope that it will be useful, ;* but WITHOUT ANY WARRANTY; without even the implied warranty of ;* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ;* Lesser General Public License for more details. ;* ;* You should have received a copy of the GNU Lesser General Public -;* License along with Libav; if not, write to the Free Software +;* License along with FFmpeg; if not, write to the Free Software ;* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA ;****************************************************************************** diff --git a/libavutil/x86/emms.asm b/libavutil/x86/emms.asm index a6851acc99..0aad34af3f 100644 --- a/libavutil/x86/emms.asm +++ b/libavutil/x86/emms.asm @@ -1,20 +1,20 @@ ;***************************************************************************** ;* Copyright (C) 2013 Martin Storsjo ;* -;* This file is part of Libav. +;* This file is part of FFmpeg. ;* -;* Libav is free software; you can redistribute it and/or +;* FFmpeg is free software; you can redistribute it and/or ;* modify it under the terms of the GNU Lesser General Public ;* License as published by the Free Software Foundation; either ;* version 2.1 of the License, or (at your option) any later version. ;* -;* Libav is distributed in the hope that it will be useful, +;* FFmpeg is distributed in the hope that it will be useful, ;* but WITHOUT ANY WARRANTY; without even the implied warranty of ;* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ;* Lesser General Public License for more details. ;* ;* You should have received a copy of the GNU Lesser General Public -;* License along with Libav; if not, write to the Free Software +;* License along with FFmpeg; if not, write to the Free Software ;* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA ;****************************************************************************** diff --git a/libavutil/x86/emms.h b/libavutil/x86/emms.h index 2ed9e5d09d..6fda6e2763 100644 --- a/libavutil/x86/emms.h +++ b/libavutil/x86/emms.h @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -21,6 +21,7 @@ #include "config.h" #include "libavutil/attributes.h" +#include "libavutil/cpu.h" void avpriv_emms_yasm(void); @@ -33,7 +34,14 @@ void avpriv_emms_yasm(void); */ static av_always_inline void emms_c(void) { - __asm__ volatile ("emms" ::: "memory"); +/* Some inlined functions may also use mmx instructions regardless of + * runtime cpuflags. With that in mind, we unconditionally empty the + * mmx state if the target cpu chosen at configure time supports it. + */ +#if !defined(__MMX__) + if(av_get_cpu_flags() & AV_CPU_FLAG_MMX) +#endif + __asm__ volatile ("emms" ::: "memory"); } #elif HAVE_MMX && HAVE_MM_EMPTY # include <mmintrin.h> diff --git a/libavutil/x86/fixed_dsp.asm b/libavutil/x86/fixed_dsp.asm new file mode 100644 index 0000000000..979dd5c334 --- /dev/null +++ b/libavutil/x86/fixed_dsp.asm @@ -0,0 +1,48 @@ +;***************************************************************************** +;* x86-optimized Float DSP functions +;* +;* Copyright 2016 James Almer +;* +;* This file is part of FFmpeg. +;* +;* FFmpeg is free software; you can redistribute it and/or +;* modify it under the terms of the GNU Lesser General Public +;* License as published by the Free Software Foundation; either +;* version 2.1 of the License, or (at your option) any later version. +;* +;* FFmpeg is distributed in the hope that it will be useful, +;* but WITHOUT ANY WARRANTY; without even the implied warranty of +;* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +;* Lesser General Public License for more details. +;* +;* You should have received a copy of the GNU Lesser General Public +;* License along with FFmpeg; if not, write to the Free Software +;* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA +;****************************************************************************** + +%include "x86util.asm" + +SECTION .text + +;----------------------------------------------------------------------------- +; void ff_butterflies_fixed(float *src0, float *src1, int len); +;----------------------------------------------------------------------------- +INIT_XMM sse2 +cglobal butterflies_fixed, 3,3,3, src0, src1, len + shl lend, 2 + add src0q, lenq + add src1q, lenq + neg lenq + +align 16 +.loop: + mova m0, [src0q + lenq] + mova m1, [src1q + lenq] + mova m2, m0 + paddd m0, m1 + psubd m2, m1 + mova [src0q + lenq], m0 + mova [src1q + lenq], m2 + add lenq, mmsize + jl .loop + RET diff --git a/libavutil/x86/fixed_dsp_init.c b/libavutil/x86/fixed_dsp_init.c new file mode 100644 index 0000000000..303a2eb922 --- /dev/null +++ b/libavutil/x86/fixed_dsp_init.c @@ -0,0 +1,35 @@ +/* + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "config.h" + +#include "libavutil/attributes.h" +#include "libavutil/cpu.h" +#include "libavutil/fixed_dsp.h" +#include "cpu.h" + +void ff_butterflies_fixed_sse2(int *src0, int *src1, int len); + +av_cold void ff_fixed_dsp_init_x86(AVFixedDSPContext *fdsp) +{ + int cpu_flags = av_get_cpu_flags(); + + if (EXTERNAL_SSE2(cpu_flags)) { + fdsp->butterflies_fixed = ff_butterflies_fixed_sse2; + } +} diff --git a/libavutil/x86/float_dsp.asm b/libavutil/x86/float_dsp.asm index d96249978a..021ff03c87 100644 --- a/libavutil/x86/float_dsp.asm +++ b/libavutil/x86/float_dsp.asm @@ -1,20 +1,22 @@ ;***************************************************************************** ;* x86-optimized Float DSP functions ;* -;* This file is part of Libav. +;* Copyright 2006 Loren Merritt ;* -;* Libav is free software; you can redistribute it and/or +;* This file is part of FFmpeg. +;* +;* FFmpeg is free software; you can redistribute it and/or ;* modify it under the terms of the GNU Lesser General Public ;* License as published by the Free Software Foundation; either ;* version 2.1 of the License, or (at your option) any later version. ;* -;* Libav is distributed in the hope that it will be useful, +;* FFmpeg is distributed in the hope that it will be useful, ;* but WITHOUT ANY WARRANTY; without even the implied warranty of ;* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ;* Lesser General Public License for more details. ;* ;* You should have received a copy of the GNU Lesser General Public -;* License along with Libav; if not, write to the Free Software +;* License along with FFmpeg; if not, write to the Free Software ;* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA ;****************************************************************************** @@ -48,8 +50,10 @@ ALIGN 16 INIT_XMM sse VECTOR_FMUL +%if HAVE_AVX_EXTERNAL INIT_YMM avx VECTOR_FMUL +%endif ;------------------------------------------------------------------------------ ; void ff_vector_fmac_scalar(float *dst, const float *src, float mul, int len) @@ -57,33 +61,48 @@ VECTOR_FMUL %macro VECTOR_FMAC_SCALAR 0 %if UNIX64 -cglobal vector_fmac_scalar, 3,3,3, dst, src, len +cglobal vector_fmac_scalar, 3,3,5, dst, src, len %else -cglobal vector_fmac_scalar, 4,4,3, dst, src, mul, len +cglobal vector_fmac_scalar, 4,4,5, dst, src, mul, len %endif %if ARCH_X86_32 VBROADCASTSS m0, mulm %else %if WIN64 - mova xmm0, xmm2 + SWAP 0, 2 %endif - shufps xmm0, xmm0, 0 + shufps xm0, xm0, 0 %if cpuflag(avx) - vinsertf128 m0, m0, xmm0, 1 + vinsertf128 m0, m0, xm0, 1 %endif %endif lea lenq, [lend*4-64] .loop: -%assign a 0 -%rep 32/mmsize - mulps m1, m0, [srcq+lenq+(a+0)*mmsize] - mulps m2, m0, [srcq+lenq+(a+1)*mmsize] - addps m1, m1, [dstq+lenq+(a+0)*mmsize] - addps m2, m2, [dstq+lenq+(a+1)*mmsize] - mova [dstq+lenq+(a+0)*mmsize], m1 - mova [dstq+lenq+(a+1)*mmsize], m2 -%assign a a+2 -%endrep +%if cpuflag(fma3) + mova m1, [dstq+lenq] + mova m2, [dstq+lenq+1*mmsize] + fmaddps m1, m0, [srcq+lenq], m1 + fmaddps m2, m0, [srcq+lenq+1*mmsize], m2 +%else ; cpuflag + mulps m1, m0, [srcq+lenq] + mulps m2, m0, [srcq+lenq+1*mmsize] +%if mmsize < 32 + mulps m3, m0, [srcq+lenq+2*mmsize] + mulps m4, m0, [srcq+lenq+3*mmsize] +%endif ; mmsize + addps m1, m1, [dstq+lenq] + addps m2, m2, [dstq+lenq+1*mmsize] +%if mmsize < 32 + addps m3, m3, [dstq+lenq+2*mmsize] + addps m4, m4, [dstq+lenq+3*mmsize] +%endif ; mmsize +%endif ; cpuflag + mova [dstq+lenq], m1 + mova [dstq+lenq+1*mmsize], m2 +%if mmsize < 32 + mova [dstq+lenq+2*mmsize], m3 + mova [dstq+lenq+3*mmsize], m4 +%endif ; mmsize sub lenq, 64 jge .loop REP_RET @@ -91,8 +110,14 @@ cglobal vector_fmac_scalar, 4,4,3, dst, src, mul, len INIT_XMM sse VECTOR_FMAC_SCALAR +%if HAVE_AVX_EXTERNAL INIT_YMM avx VECTOR_FMAC_SCALAR +%endif +%if HAVE_FMA3_EXTERNAL +INIT_YMM fma3 +VECTOR_FMAC_SCALAR +%endif ;------------------------------------------------------------------------------ ; void ff_vector_fmul_scalar(float *dst, const float *src, float mul, int len) @@ -141,16 +166,11 @@ cglobal vector_dmul_scalar, 4,4,3, dst, src, mul, len VBROADCASTSD m0, mulm %else %if WIN64 - movlhps xmm2, xmm2 -%if cpuflag(avx) - vinsertf128 ymm2, ymm2, xmm2, 1 -%endif SWAP 0, 2 -%else - movlhps xmm0, xmm0 -%if cpuflag(avx) - vinsertf128 ymm0, ymm0, xmm0, 1 %endif + movlhps xm0, xm0 +%if cpuflag(avx) + vinsertf128 ym0, ym0, xm0, 1 %endif %endif lea lenq, [lend*8-2*mmsize] @@ -172,20 +192,85 @@ VECTOR_DMUL_SCALAR %endif ;----------------------------------------------------------------------------- +; vector_fmul_window(float *dst, const float *src0, +; const float *src1, const float *win, int len); +;----------------------------------------------------------------------------- +%macro VECTOR_FMUL_WINDOW 0 +cglobal vector_fmul_window, 5, 6, 6, dst, src0, src1, win, len, len1 + shl lend, 2 + lea len1q, [lenq - mmsize] + add src0q, lenq + add dstq, lenq + add winq, lenq + neg lenq +.loop: + mova m0, [winq + lenq] + mova m4, [src0q + lenq] +%if cpuflag(sse) + mova m1, [winq + len1q] + mova m5, [src1q + len1q] + shufps m1, m1, 0x1b + shufps m5, m5, 0x1b + mova m2, m0 + mova m3, m1 + mulps m2, m4 + mulps m3, m5 + mulps m1, m4 + mulps m0, m5 + addps m2, m3 + subps m1, m0 + shufps m2, m2, 0x1b +%else + pswapd m1, [winq + len1q] + pswapd m5, [src1q + len1q] + mova m2, m0 + mova m3, m1 + pfmul m2, m4 + pfmul m3, m5 + pfmul m1, m4 + pfmul m0, m5 + pfadd m2, m3 + pfsub m1, m0 + pswapd m2, m2 +%endif + mova [dstq + lenq], m1 + mova [dstq + len1q], m2 + sub len1q, mmsize + add lenq, mmsize + jl .loop +%if mmsize == 8 + femms +%endif + REP_RET +%endmacro + +INIT_MMX 3dnowext +VECTOR_FMUL_WINDOW +INIT_XMM sse +VECTOR_FMUL_WINDOW + +;----------------------------------------------------------------------------- ; vector_fmul_add(float *dst, const float *src0, const float *src1, ; const float *src2, int len) ;----------------------------------------------------------------------------- %macro VECTOR_FMUL_ADD 0 -cglobal vector_fmul_add, 5,5,2, dst, src0, src1, src2, len +cglobal vector_fmul_add, 5,5,4, dst, src0, src1, src2, len lea lenq, [lend*4 - 2*mmsize] ALIGN 16 .loop: mova m0, [src0q + lenq] mova m1, [src0q + lenq + mmsize] +%if cpuflag(fma3) + mova m2, [src2q + lenq] + mova m3, [src2q + lenq + mmsize] + fmaddps m0, m0, [src1q + lenq], m2 + fmaddps m1, m1, [src1q + lenq + mmsize], m3 +%else mulps m0, m0, [src1q + lenq] mulps m1, m1, [src1q + lenq + mmsize] addps m0, m0, [src2q + lenq] addps m1, m1, [src2q + lenq + mmsize] +%endif mova [dstq + lenq], m0 mova [dstq + lenq + mmsize], m1 @@ -196,8 +281,14 @@ ALIGN 16 INIT_XMM sse VECTOR_FMUL_ADD +%if HAVE_AVX_EXTERNAL INIT_YMM avx VECTOR_FMUL_ADD +%endif +%if HAVE_FMA3_EXTERNAL +INIT_YMM fma3 +VECTOR_FMUL_ADD +%endif ;----------------------------------------------------------------------------- ; void vector_fmul_reverse(float *dst, const float *src0, const float *src1, @@ -233,16 +324,18 @@ ALIGN 16 INIT_XMM sse VECTOR_FMUL_REVERSE +%if HAVE_AVX_EXTERNAL INIT_YMM avx VECTOR_FMUL_REVERSE +%endif ; float scalarproduct_float_sse(const float *v1, const float *v2, int len) INIT_XMM sse cglobal scalarproduct_float, 3,3,2, v1, v2, offset + shl offsetd, 2 + add v1q, offsetq + add v2q, offsetq neg offsetq - shl offsetq, 2 - sub v1q, offsetq - sub v2q, offsetq xorps xmm0, xmm0 .loop: movaps xmm1, [v1q+offsetq] @@ -266,14 +359,9 @@ cglobal scalarproduct_float, 3,3,2, v1, v2, offset ;----------------------------------------------------------------------------- INIT_XMM sse cglobal butterflies_float, 3,3,3, src0, src1, len -%if ARCH_X86_64 - movsxd lenq, lend -%endif - test lenq, lenq - jz .end - shl lenq, 2 - lea src0q, [src0q + lenq] - lea src1q, [src1q + lenq] + shl lend, 2 + add src0q, lenq + add src1q, lenq neg lenq .loop: mova m0, [src0q + lenq] @@ -284,5 +372,4 @@ cglobal butterflies_float, 3,3,3, src0, src1, len mova [src0q + lenq], m0 add lenq, mmsize jl .loop -.end: REP_RET diff --git a/libavutil/x86/float_dsp_init.c b/libavutil/x86/float_dsp_init.c index b70433031a..c836a78e1b 100644 --- a/libavutil/x86/float_dsp_init.c +++ b/libavutil/x86/float_dsp_init.c @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -33,6 +33,8 @@ void ff_vector_fmac_scalar_sse(float *dst, const float *src, float mul, int len); void ff_vector_fmac_scalar_avx(float *dst, const float *src, float mul, int len); +void ff_vector_fmac_scalar_fma3(float *dst, const float *src, float mul, + int len); void ff_vector_fmul_scalar_sse(float *dst, const float *src, float mul, int len); @@ -42,10 +44,17 @@ void ff_vector_dmul_scalar_sse2(double *dst, const double *src, void ff_vector_dmul_scalar_avx(double *dst, const double *src, double mul, int len); +void ff_vector_fmul_window_3dnowext(float *dst, const float *src0, + const float *src1, const float *win, int len); +void ff_vector_fmul_window_sse(float *dst, const float *src0, + const float *src1, const float *win, int len); + void ff_vector_fmul_add_sse(float *dst, const float *src0, const float *src1, const float *src2, int len); void ff_vector_fmul_add_avx(float *dst, const float *src0, const float *src1, const float *src2, int len); +void ff_vector_fmul_add_fma3(float *dst, const float *src0, const float *src1, + const float *src2, int len); void ff_vector_fmul_reverse_sse(float *dst, const float *src0, const float *src1, int len); @@ -56,88 +65,18 @@ float ff_scalarproduct_float_sse(const float *v1, const float *v2, int order); void ff_butterflies_float_sse(float *src0, float *src1, int len); -#if HAVE_6REGS && HAVE_INLINE_ASM -static void vector_fmul_window_3dnowext(float *dst, const float *src0, - const float *src1, const float *win, - int len) -{ - x86_reg i = -len * 4; - x86_reg j = len * 4 - 8; - __asm__ volatile ( - "1: \n" - "pswapd (%5, %1), %%mm1 \n" - "movq (%5, %0), %%mm0 \n" - "pswapd (%4, %1), %%mm5 \n" - "movq (%3, %0), %%mm4 \n" - "movq %%mm0, %%mm2 \n" - "movq %%mm1, %%mm3 \n" - "pfmul %%mm4, %%mm2 \n" // src0[len + i] * win[len + i] - "pfmul %%mm5, %%mm3 \n" // src1[j] * win[len + j] - "pfmul %%mm4, %%mm1 \n" // src0[len + i] * win[len + j] - "pfmul %%mm5, %%mm0 \n" // src1[j] * win[len + i] - "pfadd %%mm3, %%mm2 \n" - "pfsub %%mm0, %%mm1 \n" - "pswapd %%mm2, %%mm2 \n" - "movq %%mm1, (%2, %0) \n" - "movq %%mm2, (%2, %1) \n" - "sub $8, %1 \n" - "add $8, %0 \n" - "jl 1b \n" - "femms \n" - : "+r"(i), "+r"(j) - : "r"(dst + len), "r"(src0 + len), "r"(src1), "r"(win + len) - ); -} - -static void vector_fmul_window_sse(float *dst, const float *src0, - const float *src1, const float *win, int len) -{ - x86_reg i = -len * 4; - x86_reg j = len * 4 - 16; - __asm__ volatile ( - "1: \n" - "movaps (%5, %1), %%xmm1 \n" - "movaps (%5, %0), %%xmm0 \n" - "movaps (%4, %1), %%xmm5 \n" - "movaps (%3, %0), %%xmm4 \n" - "shufps $0x1b, %%xmm1, %%xmm1 \n" - "shufps $0x1b, %%xmm5, %%xmm5 \n" - "movaps %%xmm0, %%xmm2 \n" - "movaps %%xmm1, %%xmm3 \n" - "mulps %%xmm4, %%xmm2 \n" // src0[len + i] * win[len + i] - "mulps %%xmm5, %%xmm3 \n" // src1[j] * win[len + j] - "mulps %%xmm4, %%xmm1 \n" // src0[len + i] * win[len + j] - "mulps %%xmm5, %%xmm0 \n" // src1[j] * win[len + i] - "addps %%xmm3, %%xmm2 \n" - "subps %%xmm0, %%xmm1 \n" - "shufps $0x1b, %%xmm2, %%xmm2 \n" - "movaps %%xmm1, (%2, %0) \n" - "movaps %%xmm2, (%2, %1) \n" - "sub $16, %1 \n" - "add $16, %0 \n" - "jl 1b \n" - : "+r"(i), "+r"(j) - : "r"(dst + len), "r"(src0 + len), "r"(src1), "r"(win + len) - ); -} -#endif /* HAVE_6REGS && HAVE_INLINE_ASM */ - av_cold void ff_float_dsp_init_x86(AVFloatDSPContext *fdsp) { int cpu_flags = av_get_cpu_flags(); -#if HAVE_6REGS && HAVE_INLINE_ASM - if (INLINE_AMD3DNOWEXT(cpu_flags)) { - fdsp->vector_fmul_window = vector_fmul_window_3dnowext; + if (EXTERNAL_AMD3DNOWEXT(cpu_flags)) { + fdsp->vector_fmul_window = ff_vector_fmul_window_3dnowext; } - if (INLINE_SSE(cpu_flags)) { - fdsp->vector_fmul_window = vector_fmul_window_sse; - } -#endif if (EXTERNAL_SSE(cpu_flags)) { fdsp->vector_fmul = ff_vector_fmul_sse; fdsp->vector_fmac_scalar = ff_vector_fmac_scalar_sse; fdsp->vector_fmul_scalar = ff_vector_fmul_scalar_sse; + fdsp->vector_fmul_window = ff_vector_fmul_window_sse; fdsp->vector_fmul_add = ff_vector_fmul_add_sse; fdsp->vector_fmul_reverse = ff_vector_fmul_reverse_sse; fdsp->scalarproduct_float = ff_scalarproduct_float_sse; @@ -153,4 +92,8 @@ av_cold void ff_float_dsp_init_x86(AVFloatDSPContext *fdsp) fdsp->vector_fmul_add = ff_vector_fmul_add_avx; fdsp->vector_fmul_reverse = ff_vector_fmul_reverse_avx; } + if (EXTERNAL_FMA3_FAST(cpu_flags)) { + fdsp->vector_fmac_scalar = ff_vector_fmac_scalar_fma3; + fdsp->vector_fmul_add = ff_vector_fmul_add_fma3; + } } diff --git a/libavutil/x86/intmath.h b/libavutil/x86/intmath.h new file mode 100644 index 0000000000..f58b0d08da --- /dev/null +++ b/libavutil/x86/intmath.h @@ -0,0 +1,136 @@ +/* + * Copyright (c) 2015 James Almer + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#ifndef AVUTIL_X86_INTMATH_H +#define AVUTIL_X86_INTMATH_H + +#include <stdint.h> +#include <stdlib.h> +#if HAVE_FAST_CLZ +#if defined(_MSC_VER) +#include <intrin.h> +#elif defined(__INTEL_COMPILER) +#include <immintrin.h> +#endif +#endif +#include "config.h" + +#if HAVE_FAST_CLZ +#if (defined(__INTEL_COMPILER) && (__INTEL_COMPILER>=1216)) || defined(_MSC_VER) +# if defined(__INTEL_COMPILER) +# define ff_log2(x) (_bit_scan_reverse((x)|1)) +# else +# define ff_log2 ff_log2_x86 +static av_always_inline av_const int ff_log2_x86(unsigned int v) +{ + unsigned long n; + _BitScanReverse(&n, v|1); + return n; +} +# endif +# define ff_log2_16bit av_log2 + +# define ff_ctz(v) _tzcnt_u32(v) + +# if ARCH_X86_64 +# define ff_ctzll(v) _tzcnt_u64(v) +# else +# define ff_ctzll ff_ctzll_x86 +static av_always_inline av_const int ff_ctzll_x86(long long v) +{ + return ((uint32_t)v == 0) ? _tzcnt_u32((uint32_t)(v >> 32)) + 32 : _tzcnt_u32((uint32_t)v); +} +# endif + +#endif /* __INTEL_COMPILER */ + +#endif /* HAVE_FAST_CLZ */ + +#if defined(__GNUC__) + +/* Our generic version of av_popcount is faster than GCC's built-in on + * CPUs that don't support the popcnt instruction. + */ +#if defined(__POPCNT__) + #define av_popcount __builtin_popcount +#if ARCH_X86_64 + #define av_popcount64 __builtin_popcountll +#endif + +#endif /* __POPCNT__ */ + +#if defined(__BMI2__) + +#if AV_GCC_VERSION_AT_LEAST(5,1) +#define av_mod_uintp2 __builtin_ia32_bzhi_si +#elif HAVE_INLINE_ASM +/* GCC releases before 5.1.0 have a broken bzhi builtin, so for those we + * implement it using inline assembly + */ +#define av_mod_uintp2 av_mod_uintp2_bmi2 +static av_always_inline av_const unsigned av_mod_uintp2_bmi2(unsigned a, unsigned p) +{ + if (av_builtin_constant_p(p)) + return a & ((1 << p) - 1); + else { + unsigned x; + __asm__ ("bzhi %2, %1, %0 \n\t" : "=r"(x) : "rm"(a), "r"(p)); + return x; + } +} +#endif /* AV_GCC_VERSION_AT_LEAST */ + +#endif /* __BMI2__ */ + +#if defined(__SSE2__) && !defined(__INTEL_COMPILER) + +#define av_clipd av_clipd_sse2 +static av_always_inline av_const double av_clipd_sse2(double a, double amin, double amax) +{ +#if defined(ASSERT_LEVEL) && ASSERT_LEVEL >= 2 + if (amin > amax) abort(); +#endif + __asm__ ("minsd %2, %0 \n\t" + "maxsd %1, %0 \n\t" + : "+&x"(a) : "xm"(amin), "xm"(amax)); + return a; +} + +#endif /* __SSE2__ */ + +#if defined(__SSE__) && !defined(__INTEL_COMPILER) + +#define av_clipf av_clipf_sse +static av_always_inline av_const float av_clipf_sse(float a, float amin, float amax) +{ +#if defined(ASSERT_LEVEL) && ASSERT_LEVEL >= 2 + if (amin > amax) abort(); +#endif + __asm__ ("minss %2, %0 \n\t" + "maxss %1, %0 \n\t" + : "+&x"(a) : "xm"(amin), "xm"(amax)); + return a; +} + +#endif /* __SSE__ */ + +#endif /* __GNUC__ */ + +#endif /* AVUTIL_X86_INTMATH_H */ diff --git a/libavutil/x86/intreadwrite.h b/libavutil/x86/intreadwrite.h index 635096e569..4061d19231 100644 --- a/libavutil/x86/intreadwrite.h +++ b/libavutil/x86/intreadwrite.h @@ -1,20 +1,20 @@ /* * Copyright (c) 2010 Alexander Strange <astrange@ithinksw.com> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ diff --git a/libavutil/x86/lls.asm b/libavutil/x86/lls.asm index eab85ed050..317fba6fca 100644 --- a/libavutil/x86/lls.asm +++ b/libavutil/x86/lls.asm @@ -3,20 +3,20 @@ ;* ;* Copyright (c) 2013 Loren Merritt ;* -;* This file is part of Libav. +;* This file is part of FFmpeg. ;* -;* Libav is free software; you can redistribute it and/or +;* FFmpeg is free software; you can redistribute it and/or ;* modify it under the terms of the GNU Lesser General Public ;* License as published by the Free Software Foundation; either ;* version 2.1 of the License, or (at your option) any later version. ;* -;* Libav is distributed in the hope that it will be useful, +;* FFmpeg is distributed in the hope that it will be useful, ;* but WITHOUT ANY WARRANTY; without even the implied warranty of ;* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ;* Lesser General Public License for more details. ;* ;* You should have received a copy of the GNU Lesser General Public -;* License along with Libav; if not, write to the Free Software +;* License along with FFmpeg; if not, write to the Free Software ;* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA ;****************************************************************************** @@ -125,7 +125,7 @@ cglobal update_lls, 2,5,8, ctx, var, i, j, covar2 .ret: REP_RET -INIT_YMM avx +%macro UPDATE_LLS 0 cglobal update_lls, 3,6,8, ctx, var, count, i, j, count2 %define covarq ctxq mov countd, [ctxq + LLSModel.indep_count] @@ -139,6 +139,18 @@ cglobal update_lls, 3,6,8, ctx, var, count, i, j, count2 vbroadcastsd ymm6, [varq + iq*8 + 16] vbroadcastsd ymm7, [varq + iq*8 + 24] vextractf128 xmm3, ymm1, 1 +%if cpuflag(fma3) + mova ymm0, COVAR(iq ,0) + mova xmm2, COVAR(iq+2,2) + fmaddpd ymm0, ymm1, ymm4, ymm0 + fmaddpd xmm2, xmm3, xmm6, xmm2 + fmaddpd ymm1, ymm5, ymm1, COVAR(iq ,1) + fmaddpd xmm3, xmm7, xmm3, COVAR(iq+2,3) + mova COVAR(iq ,0), ymm0 + mova COVAR(iq ,1), ymm1 + mova COVAR(iq+2,2), xmm2 + mova COVAR(iq+2,3), xmm3 +%else vmulpd ymm0, ymm1, ymm4 vmulpd ymm1, ymm1, ymm5 vmulpd xmm2, xmm3, xmm6 @@ -147,12 +159,26 @@ cglobal update_lls, 3,6,8, ctx, var, count, i, j, count2 ADDPD_MEM COVAR(iq ,1), ymm1 ADDPD_MEM COVAR(iq+2,2), xmm2 ADDPD_MEM COVAR(iq+2,3), xmm3 +%endif ; cpuflag(fma3) lea jd, [iq + 4] cmp jd, count2d jg .skip4x4 .loop4x4: ; Compute all 16 pairwise products of a 4x4 block mova ymm3, [varq + jq*8] +%if cpuflag(fma3) + mova ymm0, COVAR(jq, 0) + mova ymm1, COVAR(jq, 1) + mova ymm2, COVAR(jq, 2) + fmaddpd ymm0, ymm3, ymm4, ymm0 + fmaddpd ymm1, ymm3, ymm5, ymm1 + fmaddpd ymm2, ymm3, ymm6, ymm2 + fmaddpd ymm3, ymm7, ymm3, COVAR(jq,3) + mova COVAR(jq, 0), ymm0 + mova COVAR(jq, 1), ymm1 + mova COVAR(jq, 2), ymm2 + mova COVAR(jq, 3), ymm3 +%else vmulpd ymm0, ymm3, ymm4 vmulpd ymm1, ymm3, ymm5 vmulpd ymm2, ymm3, ymm6 @@ -161,6 +187,7 @@ cglobal update_lls, 3,6,8, ctx, var, count, i, j, count2 ADDPD_MEM COVAR(jq,1), ymm1 ADDPD_MEM COVAR(jq,2), ymm2 ADDPD_MEM COVAR(jq,3), ymm3 +%endif ; cpuflag(fma3) add jd, 4 cmp jd, count2d jle .loop4x4 @@ -168,6 +195,19 @@ cglobal update_lls, 3,6,8, ctx, var, count, i, j, count2 cmp jd, countd jg .skip2x4 mova xmm3, [varq + jq*8] +%if cpuflag(fma3) + mova xmm0, COVAR(jq, 0) + mova xmm1, COVAR(jq, 1) + mova xmm2, COVAR(jq, 2) + fmaddpd xmm0, xmm3, xmm4, xmm0 + fmaddpd xmm1, xmm3, xmm5, xmm1 + fmaddpd xmm2, xmm3, xmm6, xmm2 + fmaddpd xmm3, xmm7, xmm3, COVAR(jq,3) + mova COVAR(jq, 0), xmm0 + mova COVAR(jq, 1), xmm1 + mova COVAR(jq, 2), xmm2 + mova COVAR(jq, 3), xmm3 +%else vmulpd xmm0, xmm3, xmm4 vmulpd xmm1, xmm3, xmm5 vmulpd xmm2, xmm3, xmm6 @@ -176,6 +216,7 @@ cglobal update_lls, 3,6,8, ctx, var, count, i, j, count2 ADDPD_MEM COVAR(jq,1), xmm1 ADDPD_MEM COVAR(jq,2), xmm2 ADDPD_MEM COVAR(jq,3), xmm3 +%endif ; cpuflag(fma3) .skip2x4: add id, 4 add covarq, 4*COVAR_STRIDE @@ -186,15 +227,30 @@ cglobal update_lls, 3,6,8, ctx, var, count, i, j, count2 mov jd, id .loop2x1: vmovddup xmm0, [varq + iq*8] +%if cpuflag(fma3) + mova xmm1, [varq + jq*8] + fmaddpd xmm0, xmm1, xmm0, COVAR(jq,0) + mova COVAR(jq,0), xmm0 +%else vmulpd xmm0, [varq + jq*8] ADDPD_MEM COVAR(jq,0), xmm0 +%endif ; cpuflag(fma3) inc id add covarq, COVAR_STRIDE cmp id, countd jle .loop2x1 .ret: REP_RET +%endmacro ; UPDATE_LLS +%if HAVE_AVX_EXTERNAL +INIT_YMM avx +UPDATE_LLS +%endif +%if HAVE_FMA3_EXTERNAL +INIT_YMM fma3 +UPDATE_LLS +%endif INIT_XMM sse2 cglobal evaluate_lls, 3,4,2, ctx, var, order, i diff --git a/libavutil/x86/lls_init.c b/libavutil/x86/lls_init.c index 80cda29139..1c5dca42dc 100644 --- a/libavutil/x86/lls_init.c +++ b/libavutil/x86/lls_init.c @@ -3,29 +3,30 @@ * * Copyright (c) 2013 Loren Merritt * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ #include "libavutil/lls.h" #include "libavutil/x86/cpu.h" -void ff_update_lls_sse2(LLSModel *m, double *var); -void ff_update_lls_avx(LLSModel *m, double *var); -double ff_evaluate_lls_sse2(LLSModel *m, double *var, int order); +void ff_update_lls_sse2(LLSModel *m, const double *var); +void ff_update_lls_avx(LLSModel *m, const double *var); +void ff_update_lls_fma3(LLSModel *m, const double *var); +double ff_evaluate_lls_sse2(LLSModel *m, const double *var, int order); av_cold void ff_init_lls_x86(LLSModel *m) { @@ -38,4 +39,7 @@ av_cold void ff_init_lls_x86(LLSModel *m) if (EXTERNAL_AVX_FAST(cpu_flags)) { m->update_lls = ff_update_lls_avx; } + if (EXTERNAL_FMA3_FAST(cpu_flags)) { + m->update_lls = ff_update_lls_fma3; + } } diff --git a/libavutil/x86/pixelutils.asm b/libavutil/x86/pixelutils.asm new file mode 100644 index 0000000000..7af3007d0c --- /dev/null +++ b/libavutil/x86/pixelutils.asm @@ -0,0 +1,165 @@ +;****************************************************************************** +;* Pixel utilities SIMD +;* +;* Copyright (C) 2002-2004 Michael Niedermayer <michaelni@gmx.at> +;* Copyright (C) 2014 Clément Bœsch <u pkh me> +;* +;* This file is part of FFmpeg. +;* +;* FFmpeg is free software; you can redistribute it and/or +;* modify it under the terms of the GNU Lesser General Public +;* License as published by the Free Software Foundation; either +;* version 2.1 of the License, or (at your option) any later version. +;* +;* FFmpeg is distributed in the hope that it will be useful, +;* but WITHOUT ANY WARRANTY; without even the implied warranty of +;* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +;* Lesser General Public License for more details. +;* +;* You should have received a copy of the GNU Lesser General Public +;* License along with FFmpeg; if not, write to the Free Software +;* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA +;****************************************************************************** + +%include "x86util.asm" + +SECTION .text + +;------------------------------------------------------------------------------- +; int ff_pixelutils_sad_8x8_mmx(const uint8_t *src1, ptrdiff_t stride1, +; const uint8_t *src2, ptrdiff_t stride2); +;------------------------------------------------------------------------------- +INIT_MMX mmx +cglobal pixelutils_sad_8x8, 4,4,0, src1, stride1, src2, stride2 + pxor m7, m7 + pxor m6, m6 +%rep 4 + mova m0, [src1q] + mova m2, [src1q + stride1q] + mova m1, [src2q] + mova m3, [src2q + stride2q] + psubusb m4, m0, m1 + psubusb m5, m2, m3 + psubusb m1, m0 + psubusb m3, m2 + por m1, m4 + por m3, m5 + punpcklbw m0, m1, m7 + punpcklbw m2, m3, m7 + punpckhbw m1, m7 + punpckhbw m3, m7 + paddw m0, m1 + paddw m2, m3 + paddw m0, m2 + paddw m6, m0 + lea src1q, [src1q + 2*stride1q] + lea src2q, [src2q + 2*stride2q] +%endrep + psrlq m0, m6, 32 + paddw m6, m0 + psrlq m0, m6, 16 + paddw m6, m0 + movd eax, m6 + movzx eax, ax + RET + +;------------------------------------------------------------------------------- +; int ff_pixelutils_sad_8x8_mmxext(const uint8_t *src1, ptrdiff_t stride1, +; const uint8_t *src2, ptrdiff_t stride2); +;------------------------------------------------------------------------------- +INIT_MMX mmxext +cglobal pixelutils_sad_8x8, 4,4,0, src1, stride1, src2, stride2 + pxor m2, m2 +%rep 4 + mova m0, [src1q] + mova m1, [src1q + stride1q] + psadbw m0, [src2q] + psadbw m1, [src2q + stride2q] + paddw m2, m0 + paddw m2, m1 + lea src1q, [src1q + 2*stride1q] + lea src2q, [src2q + 2*stride2q] +%endrep + movd eax, m2 + RET + +;------------------------------------------------------------------------------- +; int ff_pixelutils_sad_16x16_mmxext(const uint8_t *src1, ptrdiff_t stride1, +; const uint8_t *src2, ptrdiff_t stride2); +;------------------------------------------------------------------------------- +INIT_MMX mmxext +cglobal pixelutils_sad_16x16, 4,4,0, src1, stride1, src2, stride2 + pxor m2, m2 +%rep 16 + mova m0, [src1q] + mova m1, [src1q + 8] + psadbw m0, [src2q] + psadbw m1, [src2q + 8] + paddw m2, m0 + paddw m2, m1 + add src1q, stride1q + add src2q, stride2q +%endrep + movd eax, m2 + RET + +;------------------------------------------------------------------------------- +; int ff_pixelutils_sad_16x16_sse(const uint8_t *src1, ptrdiff_t stride1, +; const uint8_t *src2, ptrdiff_t stride2); +;------------------------------------------------------------------------------- +INIT_XMM sse2 +cglobal pixelutils_sad_16x16, 4,4,5, src1, stride1, src2, stride2 + movu m4, [src1q] + movu m2, [src2q] + movu m1, [src1q + stride1q] + movu m3, [src2q + stride2q] + psadbw m4, m2 + psadbw m1, m3 + paddw m4, m1 +%rep 7 + lea src1q, [src1q + 2*stride1q] + lea src2q, [src2q + 2*stride2q] + movu m0, [src1q] + movu m2, [src2q] + movu m1, [src1q + stride1q] + movu m3, [src2q + stride2q] + psadbw m0, m2 + psadbw m1, m3 + paddw m4, m0 + paddw m4, m1 +%endrep + movhlps m0, m4 + paddw m4, m0 + movd eax, m4 + RET + +;------------------------------------------------------------------------------- +; int ff_pixelutils_sad_[au]_16x16_sse(const uint8_t *src1, ptrdiff_t stride1, +; const uint8_t *src2, ptrdiff_t stride2); +;------------------------------------------------------------------------------- +%macro SAD_XMM_16x16 1 +INIT_XMM sse2 +cglobal pixelutils_sad_%1_16x16, 4,4,3, src1, stride1, src2, stride2 + mov%1 m2, [src2q] + psadbw m2, [src1q] + mov%1 m1, [src2q + stride2q] + psadbw m1, [src1q + stride1q] + paddw m2, m1 +%rep 7 + lea src1q, [src1q + 2*stride1q] + lea src2q, [src2q + 2*stride2q] + mov%1 m0, [src2q] + psadbw m0, [src1q] + mov%1 m1, [src2q + stride2q] + psadbw m1, [src1q + stride1q] + paddw m2, m0 + paddw m2, m1 +%endrep + movhlps m0, m2 + paddw m2, m0 + movd eax, m2 + RET +%endmacro + +SAD_XMM_16x16 a +SAD_XMM_16x16 u diff --git a/libavutil/x86/pixelutils.h b/libavutil/x86/pixelutils.h new file mode 100644 index 0000000000..876cf46053 --- /dev/null +++ b/libavutil/x86/pixelutils.h @@ -0,0 +1,26 @@ +/* + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#ifndef AVUTIL_X86_PIXELUTILS_H +#define AVUTIL_X86_PIXELUTILS_H + +#include "libavutil/pixelutils.h" + +void ff_pixelutils_sad_init_x86(av_pixelutils_sad_fn *sad, int aligned); + +#endif /* AVUTIL_X86_PIXELUTILS_H */ diff --git a/libavutil/x86/pixelutils_init.c b/libavutil/x86/pixelutils_init.c new file mode 100644 index 0000000000..c24a533aea --- /dev/null +++ b/libavutil/x86/pixelutils_init.c @@ -0,0 +1,64 @@ +/* + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "config.h" + +#include "pixelutils.h" +#include "cpu.h" + +int ff_pixelutils_sad_8x8_mmx(const uint8_t *src1, ptrdiff_t stride1, + const uint8_t *src2, ptrdiff_t stride2); +int ff_pixelutils_sad_8x8_mmxext(const uint8_t *src1, ptrdiff_t stride1, + const uint8_t *src2, ptrdiff_t stride2); + +int ff_pixelutils_sad_16x16_mmxext(const uint8_t *src1, ptrdiff_t stride1, + const uint8_t *src2, ptrdiff_t stride2); +int ff_pixelutils_sad_16x16_sse2(const uint8_t *src1, ptrdiff_t stride1, + const uint8_t *src2, ptrdiff_t stride2); +int ff_pixelutils_sad_a_16x16_sse2(const uint8_t *src1, ptrdiff_t stride1, + const uint8_t *src2, ptrdiff_t stride2); +int ff_pixelutils_sad_u_16x16_sse2(const uint8_t *src1, ptrdiff_t stride1, + const uint8_t *src2, ptrdiff_t stride2); + +void ff_pixelutils_sad_init_x86(av_pixelutils_sad_fn *sad, int aligned) +{ + int cpu_flags = av_get_cpu_flags(); + + if (EXTERNAL_MMX(cpu_flags)) { + sad[2] = ff_pixelutils_sad_8x8_mmx; + } + + // The best way to use SSE2 would be to do 2 SADs in parallel, + // but we'd have to modify the pixelutils API to return SIMD functions. + + // It's probably not faster to shuffle data around + // to get two lines of 8 pixels into a single 16byte register, + // so just use the MMX 8x8 version even when SSE2 is available. + if (EXTERNAL_MMXEXT(cpu_flags)) { + sad[2] = ff_pixelutils_sad_8x8_mmxext; + sad[3] = ff_pixelutils_sad_16x16_mmxext; + } + + if (EXTERNAL_SSE2(cpu_flags)) { + switch (aligned) { + case 0: sad[3] = ff_pixelutils_sad_16x16_sse2; break; // src1 unaligned, src2 unaligned + case 1: sad[3] = ff_pixelutils_sad_u_16x16_sse2; break; // src1 aligned, src2 unaligned + case 2: sad[3] = ff_pixelutils_sad_a_16x16_sse2; break; // src1 aligned, src2 aligned + } + } +} diff --git a/libavutil/x86/timer.h b/libavutil/x86/timer.h index bb7c341341..4d1e88def0 100644 --- a/libavutil/x86/timer.h +++ b/libavutil/x86/timer.h @@ -1,20 +1,20 @@ /* * copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -25,6 +25,7 @@ #if HAVE_INLINE_ASM +#define FF_TIMER_UNITS "decicycles" #define AV_READ_TIME read_time static inline uint64_t read_time(void) diff --git a/libavutil/x86/w64xmmtest.h b/libavutil/x86/w64xmmtest.h index b4ce7d3daf..a4a05b0419 100644 --- a/libavutil/x86/w64xmmtest.h +++ b/libavutil/x86/w64xmmtest.h @@ -2,23 +2,26 @@ * check XMM registers for clobbers on Win64 * Copyright (c) 2008 Ramiro Polla <ramiro.polla@gmail.com> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ +#ifndef AVUTIL_X86_W64XMMTEST_H +#define AVUTIL_X86_W64XMMTEST_H + #include <inttypes.h> #include <stdint.h> #include <stdlib.h> @@ -71,3 +74,5 @@ int __real_ ## func; \ int __wrap_ ## func; \ int __wrap_ ## func + +#endif /* AVUTIL_X86_W64XMMTEST_H */ diff --git a/libavutil/x86/x86inc.asm b/libavutil/x86/x86inc.asm index 20ef7b8a19..b2e9c60195 100644 --- a/libavutil/x86/x86inc.asm +++ b/libavutil/x86/x86inc.asm @@ -798,6 +798,7 @@ BRANCH_INSTR jz, je, jnz, jne, jl, jle, jnl, jnle, jg, jge, jng, jnge, ja, jae, %assign cpuflags_atom (1<<21) %assign cpuflags_bmi1 (1<<22)|cpuflags_lzcnt %assign cpuflags_bmi2 (1<<23)|cpuflags_bmi1 +%assign cpuflags_aesni (1<<24)|cpuflags_sse42 ; Returns a boolean value expressing whether or not the specified cpuflag is enabled. %define cpuflag(x) (((((cpuflags & (cpuflags_ %+ x)) ^ (cpuflags_ %+ x)) - 1) >> 31) & 1) @@ -1096,7 +1097,7 @@ INIT_XMM ;%1 == instruction ;%2 == minimal instruction set ;%3 == 1 if float, 0 if int -;%4 == 1 if non-destructive or 4-operand (xmm, xmm, xmm, imm), 0 otherwise +;%4 == 1 if 4-operand emulation, 0 if 3-operand emulation, 255 otherwise (no emulation) ;%5 == 1 if commutative (i.e. doesn't matter which src arg is which), 0 if not ;%6+: operands %macro RUN_AVX_INSTR 6-9+ @@ -1129,14 +1130,12 @@ INIT_XMM %if __emulate_avx %xdefine __src1 %7 %xdefine __src2 %8 - %ifnidn %6, %7 - %if %0 >= 9 - CHECK_AVX_INSTR_EMU {%1 %6, %7, %8, %9}, %6, %8, %9 - %else - CHECK_AVX_INSTR_EMU {%1 %6, %7, %8}, %6, %8 - %endif - %if %5 && %4 == 0 - %ifnid %8 + %if %5 && %4 == 0 + %ifnidn %6, %7 + %ifidn %6, %8 + %xdefine __src1 %8 + %xdefine __src2 %7 + %elifnnum sizeof%8 ; 3-operand AVX instructions with a memory arg can only have it in src2, ; whereas SSE emulation prefers to have it in src1 (i.e. the mov). ; So, if the instruction is commutative with a memory arg, swap them. @@ -1144,6 +1143,13 @@ INIT_XMM %xdefine __src2 %7 %endif %endif + %endif + %ifnidn %6, __src1 + %if %0 >= 9 + CHECK_AVX_INSTR_EMU {%1 %6, %7, %8, %9}, %6, __src2, %9 + %else + CHECK_AVX_INSTR_EMU {%1 %6, %7, %8}, %6, __src2 + %endif %if __sizeofreg == 8 MOVQ %6, __src1 %elif %3 @@ -1171,9 +1177,9 @@ INIT_XMM ;%1 == instruction ;%2 == minimal instruction set ;%3 == 1 if float, 0 if int -;%4 == 1 if non-destructive or 4-operand (xmm, xmm, xmm, imm), 0 otherwise +;%4 == 1 if 4-operand emulation, 0 if 3-operand emulation, 255 otherwise (no emulation) ;%5 == 1 if commutative (i.e. doesn't matter which src arg is which), 0 if not -%macro AVX_INSTR 1-5 fnord, 0, 1, 0 +%macro AVX_INSTR 1-5 fnord, 0, 255, 0 %macro %1 1-10 fnord, fnord, fnord, fnord, %1, %2, %3, %4, %5 %ifidn %2, fnord RUN_AVX_INSTR %6, %7, %8, %9, %10, %1 @@ -1193,24 +1199,24 @@ INIT_XMM ; Non-destructive instructions are written without parameters AVX_INSTR addpd, sse2, 1, 0, 1 AVX_INSTR addps, sse, 1, 0, 1 -AVX_INSTR addsd, sse2, 1, 0, 1 -AVX_INSTR addss, sse, 1, 0, 1 +AVX_INSTR addsd, sse2, 1, 0, 0 +AVX_INSTR addss, sse, 1, 0, 0 AVX_INSTR addsubpd, sse3, 1, 0, 0 AVX_INSTR addsubps, sse3, 1, 0, 0 -AVX_INSTR aesdec, fnord, 0, 0, 0 -AVX_INSTR aesdeclast, fnord, 0, 0, 0 -AVX_INSTR aesenc, fnord, 0, 0, 0 -AVX_INSTR aesenclast, fnord, 0, 0, 0 -AVX_INSTR aesimc -AVX_INSTR aeskeygenassist +AVX_INSTR aesdec, aesni, 0, 0, 0 +AVX_INSTR aesdeclast, aesni, 0, 0, 0 +AVX_INSTR aesenc, aesni, 0, 0, 0 +AVX_INSTR aesenclast, aesni, 0, 0, 0 +AVX_INSTR aesimc, aesni +AVX_INSTR aeskeygenassist, aesni AVX_INSTR andnpd, sse2, 1, 0, 0 AVX_INSTR andnps, sse, 1, 0, 0 AVX_INSTR andpd, sse2, 1, 0, 1 AVX_INSTR andps, sse, 1, 0, 1 -AVX_INSTR blendpd, sse4, 1, 0, 0 -AVX_INSTR blendps, sse4, 1, 0, 0 -AVX_INSTR blendvpd, sse4, 1, 0, 0 -AVX_INSTR blendvps, sse4, 1, 0, 0 +AVX_INSTR blendpd, sse4, 1, 1, 0 +AVX_INSTR blendps, sse4, 1, 1, 0 +AVX_INSTR blendvpd, sse4 ; can't be emulated +AVX_INSTR blendvps, sse4 ; can't be emulated AVX_INSTR cmppd, sse2, 1, 1, 0 AVX_INSTR cmpps, sse, 1, 1, 0 AVX_INSTR cmpsd, sse2, 1, 1, 0 @@ -1224,10 +1230,10 @@ AVX_INSTR cvtpd2ps, sse2 AVX_INSTR cvtps2dq, sse2 AVX_INSTR cvtps2pd, sse2 AVX_INSTR cvtsd2si, sse2 -AVX_INSTR cvtsd2ss, sse2 -AVX_INSTR cvtsi2sd, sse2 -AVX_INSTR cvtsi2ss, sse -AVX_INSTR cvtss2sd, sse2 +AVX_INSTR cvtsd2ss, sse2, 1, 0, 0 +AVX_INSTR cvtsi2sd, sse2, 1, 0, 0 +AVX_INSTR cvtsi2ss, sse, 1, 0, 0 +AVX_INSTR cvtss2sd, sse2, 1, 0, 0 AVX_INSTR cvtss2si, sse AVX_INSTR cvttpd2dq, sse2 AVX_INSTR cvttps2dq, sse2 @@ -1250,12 +1256,12 @@ AVX_INSTR ldmxcsr, sse AVX_INSTR maskmovdqu, sse2 AVX_INSTR maxpd, sse2, 1, 0, 1 AVX_INSTR maxps, sse, 1, 0, 1 -AVX_INSTR maxsd, sse2, 1, 0, 1 -AVX_INSTR maxss, sse, 1, 0, 1 +AVX_INSTR maxsd, sse2, 1, 0, 0 +AVX_INSTR maxss, sse, 1, 0, 0 AVX_INSTR minpd, sse2, 1, 0, 1 AVX_INSTR minps, sse, 1, 0, 1 -AVX_INSTR minsd, sse2, 1, 0, 1 -AVX_INSTR minss, sse, 1, 0, 1 +AVX_INSTR minsd, sse2, 1, 0, 0 +AVX_INSTR minss, sse, 1, 0, 0 AVX_INSTR movapd, sse2 AVX_INSTR movaps, sse AVX_INSTR movd, mmx @@ -1281,11 +1287,11 @@ AVX_INSTR movsldup, sse3 AVX_INSTR movss, sse, 1, 0, 0 AVX_INSTR movupd, sse2 AVX_INSTR movups, sse -AVX_INSTR mpsadbw, sse4 +AVX_INSTR mpsadbw, sse4, 0, 1, 0 AVX_INSTR mulpd, sse2, 1, 0, 1 AVX_INSTR mulps, sse, 1, 0, 1 -AVX_INSTR mulsd, sse2, 1, 0, 1 -AVX_INSTR mulss, sse, 1, 0, 1 +AVX_INSTR mulsd, sse2, 1, 0, 0 +AVX_INSTR mulss, sse, 1, 0, 0 AVX_INSTR orpd, sse2, 1, 0, 1 AVX_INSTR orps, sse, 1, 0, 1 AVX_INSTR pabsb, ssse3 @@ -1303,14 +1309,18 @@ AVX_INSTR paddsb, mmx, 0, 0, 1 AVX_INSTR paddsw, mmx, 0, 0, 1 AVX_INSTR paddusb, mmx, 0, 0, 1 AVX_INSTR paddusw, mmx, 0, 0, 1 -AVX_INSTR palignr, ssse3 +AVX_INSTR palignr, ssse3, 0, 1, 0 AVX_INSTR pand, mmx, 0, 0, 1 AVX_INSTR pandn, mmx, 0, 0, 0 AVX_INSTR pavgb, mmx2, 0, 0, 1 AVX_INSTR pavgw, mmx2, 0, 0, 1 -AVX_INSTR pblendvb, sse4, 0, 0, 0 -AVX_INSTR pblendw, sse4 -AVX_INSTR pclmulqdq +AVX_INSTR pblendvb, sse4 ; can't be emulated +AVX_INSTR pblendw, sse4, 0, 1, 0 +AVX_INSTR pclmulqdq, fnord, 0, 1, 0 +AVX_INSTR pclmulhqhqdq, fnord, 0, 0, 0 +AVX_INSTR pclmulhqlqdq, fnord, 0, 0, 0 +AVX_INSTR pclmullqhqdq, fnord, 0, 0, 0 +AVX_INSTR pclmullqlqdq, fnord, 0, 0, 0 AVX_INSTR pcmpestri, sse42 AVX_INSTR pcmpestrm, sse42 AVX_INSTR pcmpistri, sse42 @@ -1334,10 +1344,10 @@ AVX_INSTR phminposuw, sse4 AVX_INSTR phsubw, ssse3, 0, 0, 0 AVX_INSTR phsubd, ssse3, 0, 0, 0 AVX_INSTR phsubsw, ssse3, 0, 0, 0 -AVX_INSTR pinsrb, sse4 -AVX_INSTR pinsrd, sse4 -AVX_INSTR pinsrq, sse4 -AVX_INSTR pinsrw, mmx2 +AVX_INSTR pinsrb, sse4, 0, 1, 0 +AVX_INSTR pinsrd, sse4, 0, 1, 0 +AVX_INSTR pinsrq, sse4, 0, 1, 0 +AVX_INSTR pinsrw, mmx2, 0, 1, 0 AVX_INSTR pmaddwd, mmx, 0, 0, 1 AVX_INSTR pmaddubsw, ssse3, 0, 0, 0 AVX_INSTR pmaxsb, sse4, 0, 0, 1 @@ -1409,18 +1419,18 @@ AVX_INSTR punpcklwd, mmx, 0, 0, 0 AVX_INSTR punpckldq, mmx, 0, 0, 0 AVX_INSTR punpcklqdq, sse2, 0, 0, 0 AVX_INSTR pxor, mmx, 0, 0, 1 -AVX_INSTR rcpps, sse, 1, 0, 0 +AVX_INSTR rcpps, sse AVX_INSTR rcpss, sse, 1, 0, 0 AVX_INSTR roundpd, sse4 AVX_INSTR roundps, sse4 -AVX_INSTR roundsd, sse4 -AVX_INSTR roundss, sse4 -AVX_INSTR rsqrtps, sse, 1, 0, 0 +AVX_INSTR roundsd, sse4, 1, 1, 0 +AVX_INSTR roundss, sse4, 1, 1, 0 +AVX_INSTR rsqrtps, sse AVX_INSTR rsqrtss, sse, 1, 0, 0 AVX_INSTR shufpd, sse2, 1, 1, 0 AVX_INSTR shufps, sse, 1, 1, 0 -AVX_INSTR sqrtpd, sse2, 1, 0, 0 -AVX_INSTR sqrtps, sse, 1, 0, 0 +AVX_INSTR sqrtpd, sse2 +AVX_INSTR sqrtps, sse AVX_INSTR sqrtsd, sse2, 1, 0, 0 AVX_INSTR sqrtss, sse, 1, 0, 0 AVX_INSTR stmxcsr, sse @@ -1496,7 +1506,7 @@ FMA_INSTR pmadcswd, pmaddwd, paddd v%5%6 %1, %2, %3, %4 %elifidn %1, %2 ; If %3 or %4 is a memory operand it needs to be encoded as the last operand. - %ifid %3 + %ifnum sizeof%3 v%{5}213%6 %2, %3, %4 %else v%{5}132%6 %2, %4, %3 diff --git a/libavutil/x86/x86util.asm b/libavutil/x86/x86util.asm index 9f64dd13e1..b09fa813e2 100644 --- a/libavutil/x86/x86util.asm +++ b/libavutil/x86/x86util.asm @@ -6,20 +6,20 @@ ;* Authors: Loren Merritt <lorenm@u.washington.edu> ;* Holger Lubitz <holger@lubitz.org> ;* -;* This file is part of Libav. +;* This file is part of FFmpeg. ;* -;* Libav is free software; you can redistribute it and/or +;* FFmpeg is free software; you can redistribute it and/or ;* modify it under the terms of the GNU Lesser General Public ;* License as published by the Free Software Foundation; either ;* version 2.1 of the License, or (at your option) any later version. ;* -;* Libav is distributed in the hope that it will be useful, +;* FFmpeg is distributed in the hope that it will be useful, ;* but WITHOUT ANY WARRANTY; without even the implied warranty of ;* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU ;* Lesser General Public License for more details. ;* ;* You should have received a copy of the GNU Lesser General Public -;* License along with Libav; if not, write to the Free Software +;* License along with FFmpeg; if not, write to the Free Software ;* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA ;****************************************************************************** @@ -69,6 +69,15 @@ SWAP %2, %3 %endmacro +%macro TRANSPOSE2x4x4B 5 + SBUTTERFLY bw, %1, %2, %5 + SBUTTERFLY bw, %3, %4, %5 + SBUTTERFLY wd, %1, %3, %5 + SBUTTERFLY wd, %2, %4, %5 + SBUTTERFLY dq, %1, %2, %5 + SBUTTERFLY dq, %3, %4, %5 +%endmacro + %macro TRANSPOSE2x4x4W 5 SBUTTERFLY wd, %1, %2, %5 SBUTTERFLY wd, %3, %4, %5 @@ -98,6 +107,43 @@ SWAP %5, %2, %3 %endmacro +%macro TRANSPOSE8x4D 9-11 +%if ARCH_X86_64 + SBUTTERFLY dq, %1, %2, %9 + SBUTTERFLY dq, %3, %4, %9 + SBUTTERFLY dq, %5, %6, %9 + SBUTTERFLY dq, %7, %8, %9 + SBUTTERFLY qdq, %1, %3, %9 + SBUTTERFLY qdq, %2, %4, %9 + SBUTTERFLY qdq, %5, %7, %9 + SBUTTERFLY qdq, %6, %8, %9 + SWAP %2, %5 + SWAP %4, %7 +%else +; in: m0..m7 +; out: m0..m7, unless %11 in which case m2 is in %9 +; spills into %9 and %10 + movdqa %9, m%7 + SBUTTERFLY dq, %1, %2, %7 + movdqa %10, m%2 + movdqa m%7, %9 + SBUTTERFLY dq, %3, %4, %2 + SBUTTERFLY dq, %5, %6, %2 + SBUTTERFLY dq, %7, %8, %2 + SBUTTERFLY qdq, %1, %3, %2 + movdqa %9, m%3 + movdqa m%2, %10 + SBUTTERFLY qdq, %2, %4, %3 + SBUTTERFLY qdq, %5, %7, %3 + SBUTTERFLY qdq, %6, %8, %3 + SWAP %2, %5 + SWAP %4, %7 +%if %0<11 + movdqa m%3, %9 +%endif +%endif +%endmacro + %macro TRANSPOSE8x8W 9-11 %if ARCH_X86_64 SBUTTERFLY wd, %1, %2, %9 @@ -164,13 +210,13 @@ %endif %endmacro -%macro PSIGNW_MMX 2 +%macro PSIGNW 2 +%if cpuflag(ssse3) + psignw %1, %2 +%else pxor %1, %2 psubw %1, %2 -%endmacro - -%macro PSIGNW_SSSE3 2 - psignw %1, %2 +%endif %endmacro %macro ABS1 2 @@ -273,6 +319,44 @@ %endif %endmacro +%macro HADDD 2 ; sum junk +%if sizeof%1 == 32 +%define %2 xmm%2 + vextracti128 %2, %1, 1 +%define %1 xmm%1 + paddd %1, %2 +%endif +%if mmsize >= 16 +%if cpuflag(xop) && sizeof%1 == 16 + vphadddq %1, %1 +%endif + movhlps %2, %1 + paddd %1, %2 +%endif +%if notcpuflag(xop) || sizeof%1 != 16 +%if cpuflag(mmxext) + PSHUFLW %2, %1, q0032 +%else ; mmx + mova %2, %1 + psrlq %2, 32 +%endif + paddd %1, %2 +%endif +%undef %1 +%undef %2 +%endmacro + +%macro HADDW 2 ; reg, tmp +%if cpuflag(xop) && sizeof%1 == 16 + vphaddwq %1, %1 + movhlps %2, %1 + paddd %1, %2 +%else + pmaddwd %1, [pw_1] + HADDD %1, %2 +%endif +%endmacro + %macro PALIGNR 4-5 %if cpuflag(ssse3) %if %0==5 @@ -302,11 +386,19 @@ %endif %endmacro -%macro PAVGB 2 +%macro PAVGB 2-4 %if cpuflag(mmxext) pavgb %1, %2 %elif cpuflag(3dnow) pavgusb %1, %2 +%elif cpuflag(mmx) + movu %3, %2 + por %3, %1 + pxor %1, %2 + pand %1, %4 + psrlq %1, 1 + psubb %3, %1 + SWAP %1, %3 %endif %endmacro @@ -552,7 +644,9 @@ %endmacro %macro SPLATW 2-3 0 -%if mmsize == 16 +%if cpuflag(avx2) && %3 == 0 + vpbroadcastw %1, %2 +%elif mmsize == 16 pshuflw %1, %2, (%3)*0x55 punpcklqdq %1, %1 %elif cpuflag(mmxext) @@ -683,3 +777,19 @@ addps %1, %4 %endif %endmacro + +%macro LSHIFT 2 +%if mmsize > 8 + pslldq %1, %2 +%else + psllq %1, 8*(%2) +%endif +%endmacro + +%macro RSHIFT 2 +%if mmsize > 8 + psrldq %1, %2 +%else + psrlq %1, 8*(%2) +%endif +%endmacro diff --git a/libavutil/xga_font_data.c b/libavutil/xga_font_data.c new file mode 100644 index 0000000000..3aed3142cf --- /dev/null +++ b/libavutil/xga_font_data.c @@ -0,0 +1,417 @@ +/* + * CGA/EGA/VGA ROM font data + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +/** + * @file + * CGA/EGA/VGA ROM font data + */ + +#include <stdint.h> +#include "xga_font_data.h" + +const uint8_t avpriv_cga_font[2048] = { + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x7e, 0x81, 0xa5, 0x81, 0xbd, 0x99, 0x81, 0x7e, + 0x7e, 0xff, 0xdb, 0xff, 0xc3, 0xe7, 0xff, 0x7e, 0x6c, 0xfe, 0xfe, 0xfe, 0x7c, 0x38, 0x10, 0x00, + 0x10, 0x38, 0x7c, 0xfe, 0x7c, 0x38, 0x10, 0x00, 0x38, 0x7c, 0x38, 0xfe, 0xfe, 0x7c, 0x38, 0x7c, + 0x10, 0x10, 0x38, 0x7c, 0xfe, 0x7c, 0x38, 0x7c, 0x00, 0x00, 0x18, 0x3c, 0x3c, 0x18, 0x00, 0x00, + 0xff, 0xff, 0xe7, 0xc3, 0xc3, 0xe7, 0xff, 0xff, 0x00, 0x3c, 0x66, 0x42, 0x42, 0x66, 0x3c, 0x00, + 0xff, 0xc3, 0x99, 0xbd, 0xbd, 0x99, 0xc3, 0xff, 0x0f, 0x07, 0x0f, 0x7d, 0xcc, 0xcc, 0xcc, 0x78, + 0x3c, 0x66, 0x66, 0x66, 0x3c, 0x18, 0x7e, 0x18, 0x3f, 0x33, 0x3f, 0x30, 0x30, 0x70, 0xf0, 0xe0, + 0x7f, 0x63, 0x7f, 0x63, 0x63, 0x67, 0xe6, 0xc0, 0x99, 0x5a, 0x3c, 0xe7, 0xe7, 0x3c, 0x5a, 0x99, + 0x80, 0xe0, 0xf8, 0xfe, 0xf8, 0xe0, 0x80, 0x00, 0x02, 0x0e, 0x3e, 0xfe, 0x3e, 0x0e, 0x02, 0x00, + 0x18, 0x3c, 0x7e, 0x18, 0x18, 0x7e, 0x3c, 0x18, 0x66, 0x66, 0x66, 0x66, 0x66, 0x00, 0x66, 0x00, + 0x7f, 0xdb, 0xdb, 0x7b, 0x1b, 0x1b, 0x1b, 0x00, 0x3e, 0x63, 0x38, 0x6c, 0x6c, 0x38, 0xcc, 0x78, + 0x00, 0x00, 0x00, 0x00, 0x7e, 0x7e, 0x7e, 0x00, 0x18, 0x3c, 0x7e, 0x18, 0x7e, 0x3c, 0x18, 0xff, + 0x18, 0x3c, 0x7e, 0x18, 0x18, 0x18, 0x18, 0x00, 0x18, 0x18, 0x18, 0x18, 0x7e, 0x3c, 0x18, 0x00, + 0x00, 0x18, 0x0c, 0xfe, 0x0c, 0x18, 0x00, 0x00, 0x00, 0x30, 0x60, 0xfe, 0x60, 0x30, 0x00, 0x00, + 0x00, 0x00, 0xc0, 0xc0, 0xc0, 0xfe, 0x00, 0x00, 0x00, 0x24, 0x66, 0xff, 0x66, 0x24, 0x00, 0x00, + 0x00, 0x18, 0x3c, 0x7e, 0xff, 0xff, 0x00, 0x00, 0x00, 0xff, 0xff, 0x7e, 0x3c, 0x18, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x30, 0x78, 0x78, 0x30, 0x30, 0x00, 0x30, 0x00, + 0x6c, 0x6c, 0x6c, 0x00, 0x00, 0x00, 0x00, 0x00, 0x6c, 0x6c, 0xfe, 0x6c, 0xfe, 0x6c, 0x6c, 0x00, + 0x30, 0x7c, 0xc0, 0x78, 0x0c, 0xf8, 0x30, 0x00, 0x00, 0xc6, 0xcc, 0x18, 0x30, 0x66, 0xc6, 0x00, + 0x38, 0x6c, 0x38, 0x76, 0xdc, 0xcc, 0x76, 0x00, 0x60, 0x60, 0xc0, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x18, 0x30, 0x60, 0x60, 0x60, 0x30, 0x18, 0x00, 0x60, 0x30, 0x18, 0x18, 0x18, 0x30, 0x60, 0x00, + 0x00, 0x66, 0x3c, 0xff, 0x3c, 0x66, 0x00, 0x00, 0x00, 0x30, 0x30, 0xfc, 0x30, 0x30, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x30, 0x30, 0x60, 0x00, 0x00, 0x00, 0xfc, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x30, 0x30, 0x00, 0x06, 0x0c, 0x18, 0x30, 0x60, 0xc0, 0x80, 0x00, + 0x7c, 0xc6, 0xce, 0xde, 0xf6, 0xe6, 0x7c, 0x00, 0x30, 0x70, 0x30, 0x30, 0x30, 0x30, 0xfc, 0x00, + 0x78, 0xcc, 0x0c, 0x38, 0x60, 0xcc, 0xfc, 0x00, 0x78, 0xcc, 0x0c, 0x38, 0x0c, 0xcc, 0x78, 0x00, + 0x1c, 0x3c, 0x6c, 0xcc, 0xfe, 0x0c, 0x1e, 0x00, 0xfc, 0xc0, 0xf8, 0x0c, 0x0c, 0xcc, 0x78, 0x00, + 0x38, 0x60, 0xc0, 0xf8, 0xcc, 0xcc, 0x78, 0x00, 0xfc, 0xcc, 0x0c, 0x18, 0x30, 0x30, 0x30, 0x00, + 0x78, 0xcc, 0xcc, 0x78, 0xcc, 0xcc, 0x78, 0x00, 0x78, 0xcc, 0xcc, 0x7c, 0x0c, 0x18, 0x70, 0x00, + 0x00, 0x30, 0x30, 0x00, 0x00, 0x30, 0x30, 0x00, 0x00, 0x30, 0x30, 0x00, 0x00, 0x30, 0x30, 0x60, + 0x18, 0x30, 0x60, 0xc0, 0x60, 0x30, 0x18, 0x00, 0x00, 0x00, 0xfc, 0x00, 0x00, 0xfc, 0x00, 0x00, + 0x60, 0x30, 0x18, 0x0c, 0x18, 0x30, 0x60, 0x00, 0x78, 0xcc, 0x0c, 0x18, 0x30, 0x00, 0x30, 0x00, + 0x7c, 0xc6, 0xde, 0xde, 0xde, 0xc0, 0x78, 0x00, 0x30, 0x78, 0xcc, 0xcc, 0xfc, 0xcc, 0xcc, 0x00, + 0xfc, 0x66, 0x66, 0x7c, 0x66, 0x66, 0xfc, 0x00, 0x3c, 0x66, 0xc0, 0xc0, 0xc0, 0x66, 0x3c, 0x00, + 0xf8, 0x6c, 0x66, 0x66, 0x66, 0x6c, 0xf8, 0x00, 0xfe, 0x62, 0x68, 0x78, 0x68, 0x62, 0xfe, 0x00, + 0xfe, 0x62, 0x68, 0x78, 0x68, 0x60, 0xf0, 0x00, 0x3c, 0x66, 0xc0, 0xc0, 0xce, 0x66, 0x3e, 0x00, + 0xcc, 0xcc, 0xcc, 0xfc, 0xcc, 0xcc, 0xcc, 0x00, 0x78, 0x30, 0x30, 0x30, 0x30, 0x30, 0x78, 0x00, + 0x1e, 0x0c, 0x0c, 0x0c, 0xcc, 0xcc, 0x78, 0x00, 0xe6, 0x66, 0x6c, 0x78, 0x6c, 0x66, 0xe6, 0x00, + 0xf0, 0x60, 0x60, 0x60, 0x62, 0x66, 0xfe, 0x00, 0xc6, 0xee, 0xfe, 0xfe, 0xd6, 0xc6, 0xc6, 0x00, + 0xc6, 0xe6, 0xf6, 0xde, 0xce, 0xc6, 0xc6, 0x00, 0x38, 0x6c, 0xc6, 0xc6, 0xc6, 0x6c, 0x38, 0x00, + 0xfc, 0x66, 0x66, 0x7c, 0x60, 0x60, 0xf0, 0x00, 0x78, 0xcc, 0xcc, 0xcc, 0xdc, 0x78, 0x1c, 0x00, + 0xfc, 0x66, 0x66, 0x7c, 0x6c, 0x66, 0xe6, 0x00, 0x78, 0xcc, 0xe0, 0x70, 0x1c, 0xcc, 0x78, 0x00, + 0xfc, 0xb4, 0x30, 0x30, 0x30, 0x30, 0x78, 0x00, 0xcc, 0xcc, 0xcc, 0xcc, 0xcc, 0xcc, 0xfc, 0x00, + 0xcc, 0xcc, 0xcc, 0xcc, 0xcc, 0x78, 0x30, 0x00, 0xc6, 0xc6, 0xc6, 0xd6, 0xfe, 0xee, 0xc6, 0x00, + 0xc6, 0xc6, 0x6c, 0x38, 0x38, 0x6c, 0xc6, 0x00, 0xcc, 0xcc, 0xcc, 0x78, 0x30, 0x30, 0x78, 0x00, + 0xfe, 0xc6, 0x8c, 0x18, 0x32, 0x66, 0xfe, 0x00, 0x78, 0x60, 0x60, 0x60, 0x60, 0x60, 0x78, 0x00, + 0xc0, 0x60, 0x30, 0x18, 0x0c, 0x06, 0x02, 0x00, 0x78, 0x18, 0x18, 0x18, 0x18, 0x18, 0x78, 0x00, + 0x10, 0x38, 0x6c, 0xc6, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xff, + 0x30, 0x30, 0x18, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x78, 0x0c, 0x7c, 0xcc, 0x76, 0x00, + 0xe0, 0x60, 0x60, 0x7c, 0x66, 0x66, 0xdc, 0x00, 0x00, 0x00, 0x78, 0xcc, 0xc0, 0xcc, 0x78, 0x00, + 0x1c, 0x0c, 0x0c, 0x7c, 0xcc, 0xcc, 0x76, 0x00, 0x00, 0x00, 0x78, 0xcc, 0xfc, 0xc0, 0x78, 0x00, + 0x38, 0x6c, 0x60, 0xf0, 0x60, 0x60, 0xf0, 0x00, 0x00, 0x00, 0x76, 0xcc, 0xcc, 0x7c, 0x0c, 0xf8, + 0xe0, 0x60, 0x6c, 0x76, 0x66, 0x66, 0xe6, 0x00, 0x30, 0x00, 0x70, 0x30, 0x30, 0x30, 0x78, 0x00, + 0x0c, 0x00, 0x0c, 0x0c, 0x0c, 0xcc, 0xcc, 0x78, 0xe0, 0x60, 0x66, 0x6c, 0x78, 0x6c, 0xe6, 0x00, + 0x70, 0x30, 0x30, 0x30, 0x30, 0x30, 0x78, 0x00, 0x00, 0x00, 0xcc, 0xfe, 0xfe, 0xd6, 0xc6, 0x00, + 0x00, 0x00, 0xf8, 0xcc, 0xcc, 0xcc, 0xcc, 0x00, 0x00, 0x00, 0x78, 0xcc, 0xcc, 0xcc, 0x78, 0x00, + 0x00, 0x00, 0xdc, 0x66, 0x66, 0x7c, 0x60, 0xf0, 0x00, 0x00, 0x76, 0xcc, 0xcc, 0x7c, 0x0c, 0x1e, + 0x00, 0x00, 0xdc, 0x76, 0x66, 0x60, 0xf0, 0x00, 0x00, 0x00, 0x7c, 0xc0, 0x78, 0x0c, 0xf8, 0x00, + 0x10, 0x30, 0x7c, 0x30, 0x30, 0x34, 0x18, 0x00, 0x00, 0x00, 0xcc, 0xcc, 0xcc, 0xcc, 0x76, 0x00, + 0x00, 0x00, 0xcc, 0xcc, 0xcc, 0x78, 0x30, 0x00, 0x00, 0x00, 0xc6, 0xd6, 0xfe, 0xfe, 0x6c, 0x00, + 0x00, 0x00, 0xc6, 0x6c, 0x38, 0x6c, 0xc6, 0x00, 0x00, 0x00, 0xcc, 0xcc, 0xcc, 0x7c, 0x0c, 0xf8, + 0x00, 0x00, 0xfc, 0x98, 0x30, 0x64, 0xfc, 0x00, 0x1c, 0x30, 0x30, 0xe0, 0x30, 0x30, 0x1c, 0x00, + 0x18, 0x18, 0x18, 0x00, 0x18, 0x18, 0x18, 0x00, 0xe0, 0x30, 0x30, 0x1c, 0x30, 0x30, 0xe0, 0x00, + 0x76, 0xdc, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x10, 0x38, 0x6c, 0xc6, 0xc6, 0xfe, 0x00, + 0x78, 0xcc, 0xc0, 0xcc, 0x78, 0x18, 0x0c, 0x78, 0x00, 0xcc, 0x00, 0xcc, 0xcc, 0xcc, 0x7e, 0x00, + 0x1c, 0x00, 0x78, 0xcc, 0xfc, 0xc0, 0x78, 0x00, 0x7e, 0xc3, 0x3c, 0x06, 0x3e, 0x66, 0x3f, 0x00, + 0xcc, 0x00, 0x78, 0x0c, 0x7c, 0xcc, 0x7e, 0x00, 0xe0, 0x00, 0x78, 0x0c, 0x7c, 0xcc, 0x7e, 0x00, + 0x30, 0x30, 0x78, 0x0c, 0x7c, 0xcc, 0x7e, 0x00, 0x00, 0x00, 0x78, 0xc0, 0xc0, 0x78, 0x0c, 0x38, + 0x7e, 0xc3, 0x3c, 0x66, 0x7e, 0x60, 0x3c, 0x00, 0xcc, 0x00, 0x78, 0xcc, 0xfc, 0xc0, 0x78, 0x00, + 0xe0, 0x00, 0x78, 0xcc, 0xfc, 0xc0, 0x78, 0x00, 0xcc, 0x00, 0x70, 0x30, 0x30, 0x30, 0x78, 0x00, + 0x7c, 0xc6, 0x38, 0x18, 0x18, 0x18, 0x3c, 0x00, 0xe0, 0x00, 0x70, 0x30, 0x30, 0x30, 0x78, 0x00, + 0xc6, 0x38, 0x6c, 0xc6, 0xfe, 0xc6, 0xc6, 0x00, 0x30, 0x30, 0x00, 0x78, 0xcc, 0xfc, 0xcc, 0x00, + 0x1c, 0x00, 0xfc, 0x60, 0x78, 0x60, 0xfc, 0x00, 0x00, 0x00, 0x7f, 0x0c, 0x7f, 0xcc, 0x7f, 0x00, + 0x3e, 0x6c, 0xcc, 0xfe, 0xcc, 0xcc, 0xce, 0x00, 0x78, 0xcc, 0x00, 0x78, 0xcc, 0xcc, 0x78, 0x00, + 0x00, 0xcc, 0x00, 0x78, 0xcc, 0xcc, 0x78, 0x00, 0x00, 0xe0, 0x00, 0x78, 0xcc, 0xcc, 0x78, 0x00, + 0x78, 0xcc, 0x00, 0xcc, 0xcc, 0xcc, 0x7e, 0x00, 0x00, 0xe0, 0x00, 0xcc, 0xcc, 0xcc, 0x7e, 0x00, + 0x00, 0xcc, 0x00, 0xcc, 0xcc, 0x7c, 0x0c, 0xf8, 0xc3, 0x18, 0x3c, 0x66, 0x66, 0x3c, 0x18, 0x00, + 0xcc, 0x00, 0xcc, 0xcc, 0xcc, 0xcc, 0x78, 0x00, 0x18, 0x18, 0x7e, 0xc0, 0xc0, 0x7e, 0x18, 0x18, + 0x38, 0x6c, 0x64, 0xf0, 0x60, 0xe6, 0xfc, 0x00, 0xcc, 0xcc, 0x78, 0xfc, 0x30, 0xfc, 0x30, 0x30, + 0xf8, 0xcc, 0xcc, 0xfa, 0xc6, 0xcf, 0xc6, 0xc7, 0x0e, 0x1b, 0x18, 0x3c, 0x18, 0x18, 0xd8, 0x70, + 0x1c, 0x00, 0x78, 0x0c, 0x7c, 0xcc, 0x7e, 0x00, 0x38, 0x00, 0x70, 0x30, 0x30, 0x30, 0x78, 0x00, + 0x00, 0x1c, 0x00, 0x78, 0xcc, 0xcc, 0x78, 0x00, 0x00, 0x1c, 0x00, 0xcc, 0xcc, 0xcc, 0x7e, 0x00, + 0x00, 0xf8, 0x00, 0xf8, 0xcc, 0xcc, 0xcc, 0x00, 0xfc, 0x00, 0xcc, 0xec, 0xfc, 0xdc, 0xcc, 0x00, + 0x3c, 0x6c, 0x6c, 0x3e, 0x00, 0x7e, 0x00, 0x00, 0x38, 0x6c, 0x6c, 0x38, 0x00, 0x7c, 0x00, 0x00, + 0x30, 0x00, 0x30, 0x60, 0xc0, 0xcc, 0x78, 0x00, 0x00, 0x00, 0x00, 0xfc, 0xc0, 0xc0, 0x00, 0x00, + 0x00, 0x00, 0x00, 0xfc, 0x0c, 0x0c, 0x00, 0x00, 0xc3, 0xc6, 0xcc, 0xde, 0x33, 0x66, 0xcc, 0x0f, + 0xc3, 0xc6, 0xcc, 0xdb, 0x37, 0x6f, 0xcf, 0x03, 0x18, 0x18, 0x00, 0x18, 0x18, 0x18, 0x18, 0x00, + 0x00, 0x33, 0x66, 0xcc, 0x66, 0x33, 0x00, 0x00, 0x00, 0xcc, 0x66, 0x33, 0x66, 0xcc, 0x00, 0x00, + 0x22, 0x88, 0x22, 0x88, 0x22, 0x88, 0x22, 0x88, 0x55, 0xaa, 0x55, 0xaa, 0x55, 0xaa, 0x55, 0xaa, + 0xdb, 0x77, 0xdb, 0xee, 0xdb, 0x77, 0xdb, 0xee, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, + 0x18, 0x18, 0x18, 0x18, 0xf8, 0x18, 0x18, 0x18, 0x18, 0x18, 0xf8, 0x18, 0xf8, 0x18, 0x18, 0x18, + 0x36, 0x36, 0x36, 0x36, 0xf6, 0x36, 0x36, 0x36, 0x00, 0x00, 0x00, 0x00, 0xfe, 0x36, 0x36, 0x36, + 0x00, 0x00, 0xf8, 0x18, 0xf8, 0x18, 0x18, 0x18, 0x36, 0x36, 0xf6, 0x06, 0xf6, 0x36, 0x36, 0x36, + 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x00, 0x00, 0xfe, 0x06, 0xf6, 0x36, 0x36, 0x36, + 0x36, 0x36, 0xf6, 0x06, 0xfe, 0x00, 0x00, 0x00, 0x36, 0x36, 0x36, 0x36, 0xfe, 0x00, 0x00, 0x00, + 0x18, 0x18, 0xf8, 0x18, 0xf8, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xf8, 0x18, 0x18, 0x18, + 0x18, 0x18, 0x18, 0x18, 0x1f, 0x00, 0x00, 0x00, 0x18, 0x18, 0x18, 0x18, 0xff, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0xff, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x1f, 0x18, 0x18, 0x18, + 0x00, 0x00, 0x00, 0x00, 0xff, 0x00, 0x00, 0x00, 0x18, 0x18, 0x18, 0x18, 0xff, 0x18, 0x18, 0x18, + 0x18, 0x18, 0x1f, 0x18, 0x1f, 0x18, 0x18, 0x18, 0x36, 0x36, 0x36, 0x36, 0x37, 0x36, 0x36, 0x36, + 0x36, 0x36, 0x37, 0x30, 0x3f, 0x00, 0x00, 0x00, 0x00, 0x00, 0x3f, 0x30, 0x37, 0x36, 0x36, 0x36, + 0x36, 0x36, 0xf7, 0x00, 0xff, 0x00, 0x00, 0x00, 0x00, 0x00, 0xff, 0x00, 0xf7, 0x36, 0x36, 0x36, + 0x36, 0x36, 0x37, 0x30, 0x37, 0x36, 0x36, 0x36, 0x00, 0x00, 0xff, 0x00, 0xff, 0x00, 0x00, 0x00, + 0x36, 0x36, 0xf7, 0x00, 0xf7, 0x36, 0x36, 0x36, 0x18, 0x18, 0xff, 0x00, 0xff, 0x00, 0x00, 0x00, + 0x36, 0x36, 0x36, 0x36, 0xff, 0x00, 0x00, 0x00, 0x00, 0x00, 0xff, 0x00, 0xff, 0x18, 0x18, 0x18, + 0x00, 0x00, 0x00, 0x00, 0xff, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x3f, 0x00, 0x00, 0x00, + 0x18, 0x18, 0x1f, 0x18, 0x1f, 0x00, 0x00, 0x00, 0x00, 0x00, 0x1f, 0x18, 0x1f, 0x18, 0x18, 0x18, + 0x00, 0x00, 0x00, 0x00, 0x3f, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0xff, 0x36, 0x36, 0x36, + 0x18, 0x18, 0xff, 0x18, 0xff, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0xf8, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x1f, 0x18, 0x18, 0x18, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0x00, 0x00, 0x00, 0x00, 0xff, 0xff, 0xff, 0xff, 0xf0, 0xf0, 0xf0, 0xf0, 0xf0, 0xf0, 0xf0, 0xf0, + 0x0f, 0x0f, 0x0f, 0x0f, 0x0f, 0x0f, 0x0f, 0x0f, 0xff, 0xff, 0xff, 0xff, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x76, 0xdc, 0xc8, 0xdc, 0x76, 0x00, 0x00, 0x78, 0xcc, 0xf8, 0xcc, 0xf8, 0xc0, 0xc0, + 0x00, 0xfc, 0xcc, 0xc0, 0xc0, 0xc0, 0xc0, 0x00, 0x00, 0xfe, 0x6c, 0x6c, 0x6c, 0x6c, 0x6c, 0x00, + 0xfc, 0xcc, 0x60, 0x30, 0x60, 0xcc, 0xfc, 0x00, 0x00, 0x00, 0x7e, 0xd8, 0xd8, 0xd8, 0x70, 0x00, + 0x00, 0x66, 0x66, 0x66, 0x66, 0x7c, 0x60, 0xc0, 0x00, 0x76, 0xdc, 0x18, 0x18, 0x18, 0x18, 0x00, + 0xfc, 0x30, 0x78, 0xcc, 0xcc, 0x78, 0x30, 0xfc, 0x38, 0x6c, 0xc6, 0xfe, 0xc6, 0x6c, 0x38, 0x00, + 0x38, 0x6c, 0xc6, 0xc6, 0x6c, 0x6c, 0xee, 0x00, 0x1c, 0x30, 0x18, 0x7c, 0xcc, 0xcc, 0x78, 0x00, + 0x00, 0x00, 0x7e, 0xdb, 0xdb, 0x7e, 0x00, 0x00, 0x06, 0x0c, 0x7e, 0xdb, 0xdb, 0x7e, 0x60, 0xc0, + 0x38, 0x60, 0xc0, 0xf8, 0xc0, 0x60, 0x38, 0x00, 0x78, 0xcc, 0xcc, 0xcc, 0xcc, 0xcc, 0xcc, 0x00, + 0x00, 0xfc, 0x00, 0xfc, 0x00, 0xfc, 0x00, 0x00, 0x30, 0x30, 0xfc, 0x30, 0x30, 0x00, 0xfc, 0x00, + 0x60, 0x30, 0x18, 0x30, 0x60, 0x00, 0xfc, 0x00, 0x18, 0x30, 0x60, 0x30, 0x18, 0x00, 0xfc, 0x00, + 0x0e, 0x1b, 0x1b, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0xd8, 0xd8, 0x70, + 0x30, 0x30, 0x00, 0xfc, 0x00, 0x30, 0x30, 0x00, 0x00, 0x76, 0xdc, 0x00, 0x76, 0xdc, 0x00, 0x00, + 0x38, 0x6c, 0x6c, 0x38, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x18, 0x18, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x18, 0x00, 0x00, 0x00, 0x0f, 0x0c, 0x0c, 0x0c, 0xec, 0x6c, 0x3c, 0x1c, + 0x78, 0x6c, 0x6c, 0x6c, 0x6c, 0x00, 0x00, 0x00, 0x70, 0x18, 0x30, 0x60, 0x78, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x3c, 0x3c, 0x3c, 0x3c, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, +}; + +const uint8_t avpriv_vga16_font[4096] = { + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x7e, 0x81, 0xa5, 0x81, 0x81, 0xbd, 0x99, 0x81, 0x81, 0x7e, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x7e, 0xff, 0xdb, 0xff, 0xff, 0xc3, 0xe7, 0xff, 0xff, 0x7e, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x6c, 0xfe, 0xfe, 0xfe, 0xfe, 0x7c, 0x38, 0x10, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x10, 0x38, 0x7c, 0xfe, 0x7c, 0x38, 0x10, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x18, 0x3c, 0x3c, 0xe7, 0xe7, 0xe7, 0x18, 0x18, 0x3c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x18, 0x3c, 0x7e, 0xff, 0xff, 0x7e, 0x18, 0x18, 0x3c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x18, 0x3c, 0x3c, 0x18, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xe7, 0xc3, 0xc3, 0xe7, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x3c, 0x66, 0x42, 0x42, 0x66, 0x3c, 0x00, 0x00, 0x00, 0x00, 0x00, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xc3, 0x99, 0xbd, 0xbd, 0x99, 0xc3, 0xff, 0xff, 0xff, 0xff, 0xff, + 0x00, 0x00, 0x1e, 0x0e, 0x1a, 0x32, 0x78, 0xcc, 0xcc, 0xcc, 0xcc, 0x78, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x3c, 0x66, 0x66, 0x66, 0x66, 0x3c, 0x18, 0x7e, 0x18, 0x18, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x3f, 0x33, 0x3f, 0x30, 0x30, 0x30, 0x30, 0x70, 0xf0, 0xe0, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x7f, 0x63, 0x7f, 0x63, 0x63, 0x63, 0x63, 0x67, 0xe7, 0xe6, 0xc0, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x18, 0x18, 0xdb, 0x3c, 0xe7, 0x3c, 0xdb, 0x18, 0x18, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x80, 0xc0, 0xe0, 0xf0, 0xf8, 0xfe, 0xf8, 0xf0, 0xe0, 0xc0, 0x80, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x02, 0x06, 0x0e, 0x1e, 0x3e, 0xfe, 0x3e, 0x1e, 0x0e, 0x06, 0x02, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x18, 0x3c, 0x7e, 0x18, 0x18, 0x18, 0x7e, 0x3c, 0x18, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x66, 0x66, 0x66, 0x66, 0x66, 0x66, 0x66, 0x00, 0x66, 0x66, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x7f, 0xdb, 0xdb, 0xdb, 0x7b, 0x1b, 0x1b, 0x1b, 0x1b, 0x1b, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x7c, 0xc6, 0x60, 0x38, 0x6c, 0xc6, 0xc6, 0x6c, 0x38, 0x0c, 0xc6, 0x7c, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xfe, 0xfe, 0xfe, 0xfe, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x18, 0x3c, 0x7e, 0x18, 0x18, 0x18, 0x7e, 0x3c, 0x18, 0x7e, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x18, 0x3c, 0x7e, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x7e, 0x3c, 0x18, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x18, 0x0c, 0xfe, 0x0c, 0x18, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x30, 0x60, 0xfe, 0x60, 0x30, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xc0, 0xc0, 0xc0, 0xfe, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x24, 0x66, 0xff, 0x66, 0x24, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x10, 0x38, 0x38, 0x7c, 0x7c, 0xfe, 0xfe, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0xfe, 0xfe, 0x7c, 0x7c, 0x38, 0x38, 0x10, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x18, 0x3c, 0x3c, 0x3c, 0x18, 0x18, 0x18, 0x00, 0x18, 0x18, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x66, 0x66, 0x66, 0x24, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x6c, 0x6c, 0xfe, 0x6c, 0x6c, 0x6c, 0xfe, 0x6c, 0x6c, 0x00, 0x00, 0x00, 0x00, + 0x18, 0x18, 0x7c, 0xc6, 0xc2, 0xc0, 0x7c, 0x06, 0x06, 0x86, 0xc6, 0x7c, 0x18, 0x18, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0xc2, 0xc6, 0x0c, 0x18, 0x30, 0x60, 0xc6, 0x86, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x38, 0x6c, 0x6c, 0x38, 0x76, 0xdc, 0xcc, 0xcc, 0xcc, 0x76, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x30, 0x30, 0x30, 0x60, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x0c, 0x18, 0x30, 0x30, 0x30, 0x30, 0x30, 0x30, 0x18, 0x0c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x30, 0x18, 0x0c, 0x0c, 0x0c, 0x0c, 0x0c, 0x0c, 0x18, 0x30, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x66, 0x3c, 0xff, 0x3c, 0x66, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x18, 0x18, 0x7e, 0x18, 0x18, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x18, 0x18, 0x18, 0x30, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xfe, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x18, 0x18, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x02, 0x06, 0x0c, 0x18, 0x30, 0x60, 0xc0, 0x80, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x3c, 0x66, 0xc3, 0xc3, 0xdb, 0xdb, 0xc3, 0xc3, 0x66, 0x3c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x18, 0x38, 0x78, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x7e, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x7c, 0xc6, 0x06, 0x0c, 0x18, 0x30, 0x60, 0xc0, 0xc6, 0xfe, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x7c, 0xc6, 0x06, 0x06, 0x3c, 0x06, 0x06, 0x06, 0xc6, 0x7c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x0c, 0x1c, 0x3c, 0x6c, 0xcc, 0xfe, 0x0c, 0x0c, 0x0c, 0x1e, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0xfe, 0xc0, 0xc0, 0xc0, 0xfc, 0x06, 0x06, 0x06, 0xc6, 0x7c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x38, 0x60, 0xc0, 0xc0, 0xfc, 0xc6, 0xc6, 0xc6, 0xc6, 0x7c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0xfe, 0xc6, 0x06, 0x06, 0x0c, 0x18, 0x30, 0x30, 0x30, 0x30, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x7c, 0xc6, 0xc6, 0xc6, 0x7c, 0xc6, 0xc6, 0xc6, 0xc6, 0x7c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x7c, 0xc6, 0xc6, 0xc6, 0x7e, 0x06, 0x06, 0x06, 0x0c, 0x78, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x18, 0x18, 0x00, 0x00, 0x00, 0x18, 0x18, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x18, 0x18, 0x00, 0x00, 0x00, 0x18, 0x18, 0x30, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x06, 0x0c, 0x18, 0x30, 0x60, 0x30, 0x18, 0x0c, 0x06, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x7e, 0x00, 0x00, 0x7e, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x60, 0x30, 0x18, 0x0c, 0x06, 0x0c, 0x18, 0x30, 0x60, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x7c, 0xc6, 0xc6, 0x0c, 0x18, 0x18, 0x18, 0x00, 0x18, 0x18, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x7c, 0xc6, 0xc6, 0xde, 0xde, 0xde, 0xdc, 0xc0, 0x7c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x10, 0x38, 0x6c, 0xc6, 0xc6, 0xfe, 0xc6, 0xc6, 0xc6, 0xc6, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0xfc, 0x66, 0x66, 0x66, 0x7c, 0x66, 0x66, 0x66, 0x66, 0xfc, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x3c, 0x66, 0xc2, 0xc0, 0xc0, 0xc0, 0xc0, 0xc2, 0x66, 0x3c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0xf8, 0x6c, 0x66, 0x66, 0x66, 0x66, 0x66, 0x66, 0x6c, 0xf8, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0xfe, 0x66, 0x62, 0x68, 0x78, 0x68, 0x60, 0x62, 0x66, 0xfe, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0xfe, 0x66, 0x62, 0x68, 0x78, 0x68, 0x60, 0x60, 0x60, 0xf0, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x3c, 0x66, 0xc2, 0xc0, 0xc0, 0xde, 0xc6, 0xc6, 0x66, 0x3a, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0xc6, 0xc6, 0xc6, 0xc6, 0xfe, 0xc6, 0xc6, 0xc6, 0xc6, 0xc6, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x3c, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x3c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x1e, 0x0c, 0x0c, 0x0c, 0x0c, 0x0c, 0xcc, 0xcc, 0xcc, 0x78, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0xe6, 0x66, 0x66, 0x6c, 0x78, 0x78, 0x6c, 0x66, 0x66, 0xe6, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0xf0, 0x60, 0x60, 0x60, 0x60, 0x60, 0x60, 0x62, 0x66, 0xfe, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0xc3, 0xe7, 0xff, 0xff, 0xdb, 0xc3, 0xc3, 0xc3, 0xc3, 0xc3, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0xc6, 0xe6, 0xf6, 0xfe, 0xde, 0xce, 0xc6, 0xc6, 0xc6, 0xc6, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x7c, 0xc6, 0xc6, 0xc6, 0xc6, 0xc6, 0xc6, 0xc6, 0xc6, 0x7c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0xfc, 0x66, 0x66, 0x66, 0x7c, 0x60, 0x60, 0x60, 0x60, 0xf0, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x7c, 0xc6, 0xc6, 0xc6, 0xc6, 0xc6, 0xc6, 0xd6, 0xde, 0x7c, 0x0c, 0x0e, 0x00, 0x00, + 0x00, 0x00, 0xfc, 0x66, 0x66, 0x66, 0x7c, 0x6c, 0x66, 0x66, 0x66, 0xe6, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x7c, 0xc6, 0xc6, 0x60, 0x38, 0x0c, 0x06, 0xc6, 0xc6, 0x7c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0xff, 0xdb, 0x99, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x3c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0xc6, 0xc6, 0xc6, 0xc6, 0xc6, 0xc6, 0xc6, 0xc6, 0xc6, 0x7c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0xc3, 0xc3, 0xc3, 0xc3, 0xc3, 0xc3, 0xc3, 0x66, 0x3c, 0x18, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0xc3, 0xc3, 0xc3, 0xc3, 0xc3, 0xdb, 0xdb, 0xff, 0x66, 0x66, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0xc3, 0xc3, 0x66, 0x3c, 0x18, 0x18, 0x3c, 0x66, 0xc3, 0xc3, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0xc3, 0xc3, 0xc3, 0x66, 0x3c, 0x18, 0x18, 0x18, 0x18, 0x3c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0xff, 0xc3, 0x86, 0x0c, 0x18, 0x30, 0x60, 0xc1, 0xc3, 0xff, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x3c, 0x30, 0x30, 0x30, 0x30, 0x30, 0x30, 0x30, 0x30, 0x3c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x80, 0xc0, 0xe0, 0x70, 0x38, 0x1c, 0x0e, 0x06, 0x02, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x3c, 0x0c, 0x0c, 0x0c, 0x0c, 0x0c, 0x0c, 0x0c, 0x0c, 0x3c, 0x00, 0x00, 0x00, 0x00, + 0x10, 0x38, 0x6c, 0xc6, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xff, 0x00, 0x00, + 0x30, 0x30, 0x18, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x78, 0x0c, 0x7c, 0xcc, 0xcc, 0xcc, 0x76, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0xe0, 0x60, 0x60, 0x78, 0x6c, 0x66, 0x66, 0x66, 0x66, 0x7c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x7c, 0xc6, 0xc0, 0xc0, 0xc0, 0xc6, 0x7c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x1c, 0x0c, 0x0c, 0x3c, 0x6c, 0xcc, 0xcc, 0xcc, 0xcc, 0x76, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x7c, 0xc6, 0xfe, 0xc0, 0xc0, 0xc6, 0x7c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x38, 0x6c, 0x64, 0x60, 0xf0, 0x60, 0x60, 0x60, 0x60, 0xf0, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x76, 0xcc, 0xcc, 0xcc, 0xcc, 0xcc, 0x7c, 0x0c, 0xcc, 0x78, 0x00, + 0x00, 0x00, 0xe0, 0x60, 0x60, 0x6c, 0x76, 0x66, 0x66, 0x66, 0x66, 0xe6, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x18, 0x18, 0x00, 0x38, 0x18, 0x18, 0x18, 0x18, 0x18, 0x3c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x06, 0x06, 0x00, 0x0e, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x66, 0x66, 0x3c, 0x00, + 0x00, 0x00, 0xe0, 0x60, 0x60, 0x66, 0x6c, 0x78, 0x78, 0x6c, 0x66, 0xe6, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x38, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x3c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0xe6, 0xff, 0xdb, 0xdb, 0xdb, 0xdb, 0xdb, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0xdc, 0x66, 0x66, 0x66, 0x66, 0x66, 0x66, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x7c, 0xc6, 0xc6, 0xc6, 0xc6, 0xc6, 0x7c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0xdc, 0x66, 0x66, 0x66, 0x66, 0x66, 0x7c, 0x60, 0x60, 0xf0, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x76, 0xcc, 0xcc, 0xcc, 0xcc, 0xcc, 0x7c, 0x0c, 0x0c, 0x1e, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0xdc, 0x76, 0x66, 0x60, 0x60, 0x60, 0xf0, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x7c, 0xc6, 0x60, 0x38, 0x0c, 0xc6, 0x7c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x10, 0x30, 0x30, 0xfc, 0x30, 0x30, 0x30, 0x30, 0x36, 0x1c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0xcc, 0xcc, 0xcc, 0xcc, 0xcc, 0xcc, 0x76, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0xc3, 0xc3, 0xc3, 0xc3, 0x66, 0x3c, 0x18, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0xc3, 0xc3, 0xc3, 0xdb, 0xdb, 0xff, 0x66, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0xc3, 0x66, 0x3c, 0x18, 0x3c, 0x66, 0xc3, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0xc6, 0xc6, 0xc6, 0xc6, 0xc6, 0xc6, 0x7e, 0x06, 0x0c, 0xf8, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0xfe, 0xcc, 0x18, 0x30, 0x60, 0xc6, 0xfe, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x0e, 0x18, 0x18, 0x18, 0x70, 0x18, 0x18, 0x18, 0x18, 0x0e, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x18, 0x18, 0x18, 0x18, 0x00, 0x18, 0x18, 0x18, 0x18, 0x18, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x70, 0x18, 0x18, 0x18, 0x0e, 0x18, 0x18, 0x18, 0x18, 0x70, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x76, 0xdc, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x10, 0x38, 0x6c, 0xc6, 0xc6, 0xc6, 0xfe, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x3c, 0x66, 0xc2, 0xc0, 0xc0, 0xc0, 0xc2, 0x66, 0x3c, 0x0c, 0x06, 0x7c, 0x00, 0x00, + 0x00, 0x00, 0xcc, 0x00, 0x00, 0xcc, 0xcc, 0xcc, 0xcc, 0xcc, 0xcc, 0x76, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x0c, 0x18, 0x30, 0x00, 0x7c, 0xc6, 0xfe, 0xc0, 0xc0, 0xc6, 0x7c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x10, 0x38, 0x6c, 0x00, 0x78, 0x0c, 0x7c, 0xcc, 0xcc, 0xcc, 0x76, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0xcc, 0x00, 0x00, 0x78, 0x0c, 0x7c, 0xcc, 0xcc, 0xcc, 0x76, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x60, 0x30, 0x18, 0x00, 0x78, 0x0c, 0x7c, 0xcc, 0xcc, 0xcc, 0x76, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x38, 0x6c, 0x38, 0x00, 0x78, 0x0c, 0x7c, 0xcc, 0xcc, 0xcc, 0x76, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x3c, 0x66, 0x60, 0x60, 0x66, 0x3c, 0x0c, 0x06, 0x3c, 0x00, 0x00, 0x00, + 0x00, 0x10, 0x38, 0x6c, 0x00, 0x7c, 0xc6, 0xfe, 0xc0, 0xc0, 0xc6, 0x7c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0xc6, 0x00, 0x00, 0x7c, 0xc6, 0xfe, 0xc0, 0xc0, 0xc6, 0x7c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x60, 0x30, 0x18, 0x00, 0x7c, 0xc6, 0xfe, 0xc0, 0xc0, 0xc6, 0x7c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x66, 0x00, 0x00, 0x38, 0x18, 0x18, 0x18, 0x18, 0x18, 0x3c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x18, 0x3c, 0x66, 0x00, 0x38, 0x18, 0x18, 0x18, 0x18, 0x18, 0x3c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x60, 0x30, 0x18, 0x00, 0x38, 0x18, 0x18, 0x18, 0x18, 0x18, 0x3c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0xc6, 0x00, 0x10, 0x38, 0x6c, 0xc6, 0xc6, 0xfe, 0xc6, 0xc6, 0xc6, 0x00, 0x00, 0x00, 0x00, + 0x38, 0x6c, 0x38, 0x00, 0x38, 0x6c, 0xc6, 0xc6, 0xfe, 0xc6, 0xc6, 0xc6, 0x00, 0x00, 0x00, 0x00, + 0x18, 0x30, 0x60, 0x00, 0xfe, 0x66, 0x60, 0x7c, 0x60, 0x60, 0x66, 0xfe, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x6e, 0x3b, 0x1b, 0x7e, 0xd8, 0xdc, 0x77, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x3e, 0x6c, 0xcc, 0xcc, 0xfe, 0xcc, 0xcc, 0xcc, 0xcc, 0xce, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x10, 0x38, 0x6c, 0x00, 0x7c, 0xc6, 0xc6, 0xc6, 0xc6, 0xc6, 0x7c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0xc6, 0x00, 0x00, 0x7c, 0xc6, 0xc6, 0xc6, 0xc6, 0xc6, 0x7c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x60, 0x30, 0x18, 0x00, 0x7c, 0xc6, 0xc6, 0xc6, 0xc6, 0xc6, 0x7c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x30, 0x78, 0xcc, 0x00, 0xcc, 0xcc, 0xcc, 0xcc, 0xcc, 0xcc, 0x76, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x60, 0x30, 0x18, 0x00, 0xcc, 0xcc, 0xcc, 0xcc, 0xcc, 0xcc, 0x76, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0xc6, 0x00, 0x00, 0xc6, 0xc6, 0xc6, 0xc6, 0xc6, 0xc6, 0x7e, 0x06, 0x0c, 0x78, 0x00, + 0x00, 0xc6, 0x00, 0x7c, 0xc6, 0xc6, 0xc6, 0xc6, 0xc6, 0xc6, 0xc6, 0x7c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0xc6, 0x00, 0xc6, 0xc6, 0xc6, 0xc6, 0xc6, 0xc6, 0xc6, 0xc6, 0x7c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x18, 0x18, 0x7e, 0xc3, 0xc0, 0xc0, 0xc0, 0xc3, 0x7e, 0x18, 0x18, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x38, 0x6c, 0x64, 0x60, 0xf0, 0x60, 0x60, 0x60, 0x60, 0xe6, 0xfc, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0xc3, 0x66, 0x3c, 0x18, 0xff, 0x18, 0xff, 0x18, 0x18, 0x18, 0x00, 0x00, 0x00, 0x00, + 0x00, 0xfc, 0x66, 0x66, 0x7c, 0x62, 0x66, 0x6f, 0x66, 0x66, 0x66, 0xf3, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x0e, 0x1b, 0x18, 0x18, 0x18, 0x7e, 0x18, 0x18, 0x18, 0x18, 0x18, 0xd8, 0x70, 0x00, 0x00, + 0x00, 0x18, 0x30, 0x60, 0x00, 0x78, 0x0c, 0x7c, 0xcc, 0xcc, 0xcc, 0x76, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x0c, 0x18, 0x30, 0x00, 0x38, 0x18, 0x18, 0x18, 0x18, 0x18, 0x3c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x18, 0x30, 0x60, 0x00, 0x7c, 0xc6, 0xc6, 0xc6, 0xc6, 0xc6, 0x7c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x18, 0x30, 0x60, 0x00, 0xcc, 0xcc, 0xcc, 0xcc, 0xcc, 0xcc, 0x76, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x76, 0xdc, 0x00, 0xdc, 0x66, 0x66, 0x66, 0x66, 0x66, 0x66, 0x00, 0x00, 0x00, 0x00, + 0x76, 0xdc, 0x00, 0xc6, 0xe6, 0xf6, 0xfe, 0xde, 0xce, 0xc6, 0xc6, 0xc6, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x3c, 0x6c, 0x6c, 0x3e, 0x00, 0x7e, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x38, 0x6c, 0x6c, 0x38, 0x00, 0x7c, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x30, 0x30, 0x00, 0x30, 0x30, 0x60, 0xc0, 0xc6, 0xc6, 0x7c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xfe, 0xc0, 0xc0, 0xc0, 0xc0, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xfe, 0x06, 0x06, 0x06, 0x06, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0xc0, 0xc0, 0xc2, 0xc6, 0xcc, 0x18, 0x30, 0x60, 0xce, 0x9b, 0x06, 0x0c, 0x1f, 0x00, 0x00, + 0x00, 0xc0, 0xc0, 0xc2, 0xc6, 0xcc, 0x18, 0x30, 0x66, 0xce, 0x96, 0x3e, 0x06, 0x06, 0x00, 0x00, + 0x00, 0x00, 0x18, 0x18, 0x00, 0x18, 0x18, 0x18, 0x3c, 0x3c, 0x3c, 0x18, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x36, 0x6c, 0xd8, 0x6c, 0x36, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0xd8, 0x6c, 0x36, 0x6c, 0xd8, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x11, 0x44, 0x11, 0x44, 0x11, 0x44, 0x11, 0x44, 0x11, 0x44, 0x11, 0x44, 0x11, 0x44, 0x11, 0x44, + 0x55, 0xaa, 0x55, 0xaa, 0x55, 0xaa, 0x55, 0xaa, 0x55, 0xaa, 0x55, 0xaa, 0x55, 0xaa, 0x55, 0xaa, + 0xdd, 0x77, 0xdd, 0x77, 0xdd, 0x77, 0xdd, 0x77, 0xdd, 0x77, 0xdd, 0x77, 0xdd, 0x77, 0xdd, 0x77, + 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, + 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0xf8, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, + 0x18, 0x18, 0x18, 0x18, 0x18, 0xf8, 0x18, 0xf8, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, + 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0xf6, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xfe, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, + 0x00, 0x00, 0x00, 0x00, 0x00, 0xf8, 0x18, 0xf8, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, + 0x36, 0x36, 0x36, 0x36, 0x36, 0xf6, 0x06, 0xf6, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, + 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, + 0x00, 0x00, 0x00, 0x00, 0x00, 0xfe, 0x06, 0xf6, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, + 0x36, 0x36, 0x36, 0x36, 0x36, 0xf6, 0x06, 0xfe, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0xfe, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x18, 0x18, 0x18, 0x18, 0x18, 0xf8, 0x18, 0xf8, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xf8, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, + 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x1f, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0xff, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xff, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, + 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x1f, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xff, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0xff, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, + 0x18, 0x18, 0x18, 0x18, 0x18, 0x1f, 0x18, 0x1f, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, + 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x37, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, + 0x36, 0x36, 0x36, 0x36, 0x36, 0x37, 0x30, 0x3f, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x3f, 0x30, 0x37, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, + 0x36, 0x36, 0x36, 0x36, 0x36, 0xf7, 0x00, 0xff, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0xff, 0x00, 0xf7, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, + 0x36, 0x36, 0x36, 0x36, 0x36, 0x37, 0x30, 0x37, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, + 0x00, 0x00, 0x00, 0x00, 0x00, 0xff, 0x00, 0xff, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x36, 0x36, 0x36, 0x36, 0x36, 0xf7, 0x00, 0xf7, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, + 0x18, 0x18, 0x18, 0x18, 0x18, 0xff, 0x00, 0xff, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0xff, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0xff, 0x00, 0xff, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xff, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, + 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x3f, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x18, 0x18, 0x18, 0x18, 0x18, 0x1f, 0x18, 0x1f, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x1f, 0x18, 0x1f, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x3f, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, + 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0xff, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, 0x36, + 0x18, 0x18, 0x18, 0x18, 0x18, 0xff, 0x18, 0xff, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, + 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0xf8, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x1f, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xf0, 0xf0, 0xf0, 0xf0, 0xf0, 0xf0, 0xf0, 0xf0, 0xf0, 0xf0, 0xf0, 0xf0, 0xf0, 0xf0, 0xf0, 0xf0, + 0x0f, 0x0f, 0x0f, 0x0f, 0x0f, 0x0f, 0x0f, 0x0f, 0x0f, 0x0f, 0x0f, 0x0f, 0x0f, 0x0f, 0x0f, 0x0f, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x76, 0xdc, 0xd8, 0xd8, 0xd8, 0xdc, 0x76, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x78, 0xcc, 0xcc, 0xcc, 0xd8, 0xcc, 0xc6, 0xc6, 0xc6, 0xcc, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0xfe, 0xc6, 0xc6, 0xc0, 0xc0, 0xc0, 0xc0, 0xc0, 0xc0, 0xc0, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0xfe, 0x6c, 0x6c, 0x6c, 0x6c, 0x6c, 0x6c, 0x6c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0xfe, 0xc6, 0x60, 0x30, 0x18, 0x30, 0x60, 0xc6, 0xfe, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x7e, 0xd8, 0xd8, 0xd8, 0xd8, 0xd8, 0x70, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x66, 0x66, 0x66, 0x66, 0x66, 0x7c, 0x60, 0x60, 0xc0, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x76, 0xdc, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x7e, 0x18, 0x3c, 0x66, 0x66, 0x66, 0x3c, 0x18, 0x7e, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x38, 0x6c, 0xc6, 0xc6, 0xfe, 0xc6, 0xc6, 0x6c, 0x38, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x38, 0x6c, 0xc6, 0xc6, 0xc6, 0x6c, 0x6c, 0x6c, 0x6c, 0xee, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x1e, 0x30, 0x18, 0x0c, 0x3e, 0x66, 0x66, 0x66, 0x66, 0x3c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x7e, 0xdb, 0xdb, 0xdb, 0x7e, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x03, 0x06, 0x7e, 0xdb, 0xdb, 0xf3, 0x7e, 0x60, 0xc0, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x1c, 0x30, 0x60, 0x60, 0x7c, 0x60, 0x60, 0x60, 0x30, 0x1c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x7c, 0xc6, 0xc6, 0xc6, 0xc6, 0xc6, 0xc6, 0xc6, 0xc6, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0xfe, 0x00, 0x00, 0xfe, 0x00, 0x00, 0xfe, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x18, 0x18, 0x7e, 0x18, 0x18, 0x00, 0x00, 0xff, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x30, 0x18, 0x0c, 0x06, 0x0c, 0x18, 0x30, 0x00, 0x7e, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x0c, 0x18, 0x30, 0x60, 0x30, 0x18, 0x0c, 0x00, 0x7e, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x0e, 0x1b, 0x1b, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, + 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0x18, 0xd8, 0xd8, 0xd8, 0x70, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x18, 0x18, 0x00, 0x7e, 0x00, 0x18, 0x18, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x76, 0xdc, 0x00, 0x76, 0xdc, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x38, 0x6c, 0x6c, 0x38, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x18, 0x18, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x18, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x0f, 0x0c, 0x0c, 0x0c, 0x0c, 0x0c, 0xec, 0x6c, 0x6c, 0x3c, 0x1c, 0x00, 0x00, 0x00, 0x00, + 0x00, 0xd8, 0x6c, 0x6c, 0x6c, 0x6c, 0x6c, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x70, 0xd8, 0x30, 0x60, 0xc8, 0xf8, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x7c, 0x7c, 0x7c, 0x7c, 0x7c, 0x7c, 0x7c, 0x00, 0x00, 0x00, 0x00, 0x00, + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00 +}; diff --git a/libavutil/xga_font_data.h b/libavutil/xga_font_data.h new file mode 100644 index 0000000000..5e40f542e4 --- /dev/null +++ b/libavutil/xga_font_data.h @@ -0,0 +1,35 @@ +/* + * CGA/EGA/VGA ROM font data + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +/** + * @file + * CGA/EGA/VGA ROM font data + */ + +#ifndef AVUTIL_XGA_FONT_DATA_H +#define AVUTIL_XGA_FONT_DATA_H + +#include <stdint.h> +#include "internal.h" + +extern av_export const uint8_t avpriv_cga_font[2048]; +extern av_export const uint8_t avpriv_vga16_font[4096]; + +#endif /* AVUTIL_XGA_FONT_DATA_H */ diff --git a/libavutil/xtea-test.c b/libavutil/xtea-test.c index f81cfdc3e8..e8f39150d9 100644 --- a/libavutil/xtea-test.c +++ b/libavutil/xtea-test.c @@ -1,18 +1,18 @@ /* - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -83,7 +83,7 @@ int main(void) AVXTEA ctx; uint8_t buf[16], iv[8]; int i, j; - const uint8_t src[32] = "HelloWorldHelloWorldHelloWorld"; + static const uint8_t src[32] = "HelloWorldHelloWorldHelloWorld"; uint8_t ct[32]; uint8_t pl[32]; @@ -114,6 +114,7 @@ int main(void) memcpy(iv, "HALLO123", 8); test_xtea(&ctx, ct, ct, src, 4, iv, 1, "CBC inplace decryption", av_xtea_crypt); } + printf("Test encryption/decryption success.\n"); return 0; diff --git a/libavutil/xtea.c b/libavutil/xtea.c index 3e222365b3..f7892af9fb 100644 --- a/libavutil/xtea.c +++ b/libavutil/xtea.c @@ -4,20 +4,20 @@ * * loosely based on the implementation of David Wheeler and Roger Needham * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -34,12 +34,6 @@ #include "mem.h" #include "xtea.h" -#if !FF_API_CRYPTO_CONTEXT -struct AVXTEA { - uint32_t key[16]; -}; -#endif - AVXTEA *av_xtea_alloc(void) { return av_mallocz(sizeof(struct AVXTEA)); @@ -65,32 +59,116 @@ static void xtea_crypt_ecb(AVXTEA *ctx, uint8_t *dst, const uint8_t *src, int decrypt, uint8_t *iv) { uint32_t v0, v1; - int i; +#if !CONFIG_SMALL + uint32_t k0 = ctx->key[0]; + uint32_t k1 = ctx->key[1]; + uint32_t k2 = ctx->key[2]; + uint32_t k3 = ctx->key[3]; +#endif v0 = AV_RB32(src); v1 = AV_RB32(src + 4); if (decrypt) { - uint32_t delta = 0x9E3779B9, sum = delta * 32; +#if CONFIG_SMALL + int i; + uint32_t delta = 0x9E3779B9U, sum = delta * 32; for (i = 0; i < 32; i++) { v1 -= (((v0 << 4) ^ (v0 >> 5)) + v0) ^ (sum + ctx->key[(sum >> 11) & 3]); sum -= delta; v0 -= (((v1 << 4) ^ (v1 >> 5)) + v1) ^ (sum + ctx->key[sum & 3]); } +#else +#define DSTEP(SUM, K0, K1) \ + v1 -= (((v0 << 4) ^ (v0 >> 5)) + v0) ^ (SUM + K0); \ + v0 -= (((v1 << 4) ^ (v1 >> 5)) + v1) ^ (SUM - 0x9E3779B9U + K1) + + DSTEP(0xC6EF3720U, k2, k3); + DSTEP(0x28B7BD67U, k3, k2); + DSTEP(0x8A8043AEU, k0, k1); + DSTEP(0xEC48C9F5U, k1, k0); + DSTEP(0x4E11503CU, k2, k3); + DSTEP(0xAFD9D683U, k2, k2); + DSTEP(0x11A25CCAU, k3, k1); + DSTEP(0x736AE311U, k0, k0); + DSTEP(0xD5336958U, k1, k3); + DSTEP(0x36FBEF9FU, k1, k2); + DSTEP(0x98C475E6U, k2, k1); + DSTEP(0xFA8CFC2DU, k3, k0); + DSTEP(0x5C558274U, k0, k3); + DSTEP(0xBE1E08BBU, k1, k2); + DSTEP(0x1FE68F02U, k1, k1); + DSTEP(0x81AF1549U, k2, k0); + DSTEP(0xE3779B90U, k3, k3); + DSTEP(0x454021D7U, k0, k2); + DSTEP(0xA708A81EU, k1, k1); + DSTEP(0x08D12E65U, k1, k0); + DSTEP(0x6A99B4ACU, k2, k3); + DSTEP(0xCC623AF3U, k3, k2); + DSTEP(0x2E2AC13AU, k0, k1); + DSTEP(0x8FF34781U, k0, k0); + DSTEP(0xF1BBCDC8U, k1, k3); + DSTEP(0x5384540FU, k2, k2); + DSTEP(0xB54CDA56U, k3, k1); + DSTEP(0x1715609DU, k0, k0); + DSTEP(0x78DDE6E4U, k0, k3); + DSTEP(0xDAA66D2BU, k1, k2); + DSTEP(0x3C6EF372U, k2, k1); + DSTEP(0x9E3779B9U, k3, k0); +#endif if (iv) { v0 ^= AV_RB32(iv); v1 ^= AV_RB32(iv + 4); memcpy(iv, src, 8); } } else { - uint32_t sum = 0, delta = 0x9E3779B9; +#if CONFIG_SMALL + int i; + uint32_t sum = 0, delta = 0x9E3779B9U; for (i = 0; i < 32; i++) { v0 += (((v1 << 4) ^ (v1 >> 5)) + v1) ^ (sum + ctx->key[sum & 3]); sum += delta; v1 += (((v0 << 4) ^ (v0 >> 5)) + v0) ^ (sum + ctx->key[(sum >> 11) & 3]); } +#else +#define ESTEP(SUM, K0, K1) \ + v0 += (((v1 << 4) ^ (v1 >> 5)) + v1) ^ (SUM + K0);\ + v1 += (((v0 << 4) ^ (v0 >> 5)) + v0) ^ (SUM + 0x9E3779B9U + K1) + ESTEP(0x00000000U, k0, k3); + ESTEP(0x9E3779B9U, k1, k2); + ESTEP(0x3C6EF372U, k2, k1); + ESTEP(0xDAA66D2BU, k3, k0); + ESTEP(0x78DDE6E4U, k0, k0); + ESTEP(0x1715609DU, k1, k3); + ESTEP(0xB54CDA56U, k2, k2); + ESTEP(0x5384540FU, k3, k1); + ESTEP(0xF1BBCDC8U, k0, k0); + ESTEP(0x8FF34781U, k1, k0); + ESTEP(0x2E2AC13AU, k2, k3); + ESTEP(0xCC623AF3U, k3, k2); + ESTEP(0x6A99B4ACU, k0, k1); + ESTEP(0x08D12E65U, k1, k1); + ESTEP(0xA708A81EU, k2, k0); + ESTEP(0x454021D7U, k3, k3); + ESTEP(0xE3779B90U, k0, k2); + ESTEP(0x81AF1549U, k1, k1); + ESTEP(0x1FE68F02U, k2, k1); + ESTEP(0xBE1E08BBU, k3, k0); + ESTEP(0x5C558274U, k0, k3); + ESTEP(0xFA8CFC2DU, k1, k2); + ESTEP(0x98C475E6U, k2, k1); + ESTEP(0x36FBEF9FU, k3, k1); + ESTEP(0xD5336958U, k0, k0); + ESTEP(0x736AE311U, k1, k3); + ESTEP(0x11A25CCAU, k2, k2); + ESTEP(0xAFD9D683U, k3, k2); + ESTEP(0x4E11503CU, k0, k1); + ESTEP(0xEC48C9F5U, k1, k0); + ESTEP(0x8A8043AEU, k2, k3); + ESTEP(0x28B7BD67U, k3, k2); +#endif } AV_WB32(dst, v0); diff --git a/libavutil/xtea.h b/libavutil/xtea.h index 78b1f378db..735427c109 100644 --- a/libavutil/xtea.h +++ b/libavutil/xtea.h @@ -1,20 +1,21 @@ /* * A 32-bit implementation of the XTEA algorithm + * Copyright (c) 2012 Samuel Pitoiset * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -22,7 +23,6 @@ #define AVUTIL_XTEA_H #include <stdint.h> -#include "version.h" /** * @file @@ -32,13 +32,9 @@ * @{ */ -#if FF_API_CRYPTO_CONTEXT typedef struct AVXTEA { uint32_t key[16]; } AVXTEA; -#else -typedef struct AVXTEA AVXTEA; -#endif /** * Allocate an AVXTEA context. |