diff --git a/libavcodec/aarch64/Makefile b/libavcodec/aarch64/Makefile index 00f93bf59f..52da7036f3 100644 --- a/libavcodec/aarch64/Makefile +++ b/libavcodec/aarch64/Makefile @@ -6,6 +6,7 @@ OBJS-$(CONFIG_H264DSP) += aarch64/h264dsp_init_aarch64.o OBJS-$(CONFIG_H264PRED) += aarch64/h264pred_init.o OBJS-$(CONFIG_H264QPEL) += aarch64/h264qpel_init_aarch64.o OBJS-$(CONFIG_HPELDSP) += aarch64/hpeldsp_init_aarch64.o +OBJS-$(CONFIG_IDCTDSP) += aarch64/idctdsp_init_aarch64.o OBJS-$(CONFIG_MPEGAUDIODSP) += aarch64/mpegaudiodsp_init.o OBJS-$(CONFIG_NEON_CLOBBER_TEST) += aarch64/neontest.o OBJS-$(CONFIG_VIDEODSP) += aarch64/videodsp_init.o @@ -21,6 +22,7 @@ OBJS-$(CONFIG_VC1DSP) += aarch64/vc1dsp_init_aarch64.o OBJS-$(CONFIG_VORBIS_DECODER) += aarch64/vorbisdsp_init.o OBJS-$(CONFIG_VP9_DECODER) += aarch64/vp9dsp_init_10bpp_aarch64.o \ aarch64/vp9dsp_init_12bpp_aarch64.o \ + aarch64/vp9mc_aarch64.o \ aarch64/vp9dsp_init_aarch64.o # ARMv8 optimizations @@ -41,8 +43,7 @@ NEON-OBJS-$(CONFIG_H264PRED) += aarch64/h264pred_neon.o NEON-OBJS-$(CONFIG_H264QPEL) += aarch64/h264qpel_neon.o \ aarch64/hpeldsp_neon.o NEON-OBJS-$(CONFIG_HPELDSP) += aarch64/hpeldsp_neon.o -NEON-OBJS-$(CONFIG_IDCTDSP) += aarch64/idctdsp_init_aarch64.o \ - aarch64/simple_idct_neon.o +NEON-OBJS-$(CONFIG_IDCTDSP) += aarch64/simple_idct_neon.o NEON-OBJS-$(CONFIG_MDCT) += aarch64/mdct_neon.o NEON-OBJS-$(CONFIG_MPEGAUDIODSP) += aarch64/mpegaudiodsp_neon.o NEON-OBJS-$(CONFIG_VP8DSP) += aarch64/vp8dsp_neon.o diff --git a/libavcodec/aarch64/idctdsp_init_aarch64.c b/libavcodec/aarch64/idctdsp_init_aarch64.c index 0406e60830..742a3372e3 100644 --- a/libavcodec/aarch64/idctdsp_init_aarch64.c +++ b/libavcodec/aarch64/idctdsp_init_aarch64.c @@ -21,6 +21,8 @@ */ #include "libavutil/attributes.h" +#include "libavutil/cpu.h" +#include "libavutil/arm/cpu.h" #include "libavcodec/avcodec.h" #include "libavcodec/idctdsp.h" #include "idct.h" @@ -28,7 +30,9 @@ av_cold void ff_idctdsp_init_aarch64(IDCTDSPContext *c, AVCodecContext *avctx, unsigned high_bit_depth) { - if (!avctx->lowres && !high_bit_depth) { + int cpu_flags = av_get_cpu_flags(); + + if (have_neon(cpu_flags) && !avctx->lowres && !high_bit_depth) { if (avctx->idct_algo == FF_IDCT_AUTO || avctx->idct_algo == FF_IDCT_SIMPLEAUTO || avctx->idct_algo == FF_IDCT_SIMPLENEON) { diff --git a/libavcodec/aarch64/vp9mc_16bpp_neon.S b/libavcodec/aarch64/vp9mc_16bpp_neon.S index cac6428709..53b372c262 100644 --- a/libavcodec/aarch64/vp9mc_16bpp_neon.S +++ b/libavcodec/aarch64/vp9mc_16bpp_neon.S @@ -25,31 +25,6 @@ // const uint8_t *ref, ptrdiff_t ref_stride, // int h, int mx, int my); -function ff_vp9_copy128_aarch64, export=1 -1: - ldp x5, x6, [x2] - ldp x7, x8, [x2, #16] - stp x5, x6, [x0] - ldp x9, x10, [x2, #32] - stp x7, x8, [x0, #16] - subs w4, w4, #1 - ldp x11, x12, [x2, #48] - stp x9, x10, [x0, #32] - stp x11, x12, [x0, #48] - ldp x5, x6, [x2, #64] - ldp x7, x8, [x2, #80] - stp x5, x6, [x0, #64] - ldp x9, x10, [x2, #96] - stp x7, x8, [x0, #80] - ldp x11, x12, [x2, #112] - stp x9, x10, [x0, #96] - stp x11, x12, [x0, #112] - add x2, x2, x3 - add x0, x0, x1 - b.ne 1b - ret -endfunc - function ff_vp9_avg64_16_neon, export=1 mov x5, x0 sub x1, x1, #64 diff --git a/libavcodec/aarch64/vp9mc_aarch64.S b/libavcodec/aarch64/vp9mc_aarch64.S new file mode 100644 index 0000000000..f17a8cf04a --- /dev/null +++ b/libavcodec/aarch64/vp9mc_aarch64.S @@ -0,0 +1,81 @@ +/* + * Copyright (c) 2016 Google Inc. + * + * 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 "libavutil/aarch64/asm.S" + +// All public functions in this file have the following signature: +// typedef void (*vp9_mc_func)(uint8_t *dst, ptrdiff_t dst_stride, +// const uint8_t *ref, ptrdiff_t ref_stride, +// int h, int mx, int my); + +function ff_vp9_copy128_aarch64, export=1 +1: + ldp x5, x6, [x2] + ldp x7, x8, [x2, #16] + stp x5, x6, [x0] + ldp x9, x10, [x2, #32] + stp x7, x8, [x0, #16] + subs w4, w4, #1 + ldp x11, x12, [x2, #48] + stp x9, x10, [x0, #32] + stp x11, x12, [x0, #48] + ldp x5, x6, [x2, #64] + ldp x7, x8, [x2, #80] + stp x5, x6, [x0, #64] + ldp x9, x10, [x2, #96] + stp x7, x8, [x0, #80] + ldp x11, x12, [x2, #112] + stp x9, x10, [x0, #96] + stp x11, x12, [x0, #112] + add x2, x2, x3 + add x0, x0, x1 + b.ne 1b + ret +endfunc + +function ff_vp9_copy64_aarch64, export=1 +1: + ldp x5, x6, [x2] + ldp x7, x8, [x2, #16] + stp x5, x6, [x0] + ldp x9, x10, [x2, #32] + stp x7, x8, [x0, #16] + subs w4, w4, #1 + ldp x11, x12, [x2, #48] + stp x9, x10, [x0, #32] + stp x11, x12, [x0, #48] + add x2, x2, x3 + add x0, x0, x1 + b.ne 1b + ret +endfunc + +function ff_vp9_copy32_aarch64, export=1 +1: + ldp x5, x6, [x2] + ldp x7, x8, [x2, #16] + stp x5, x6, [x0] + subs w4, w4, #1 + stp x7, x8, [x0, #16] + add x2, x2, x3 + add x0, x0, x1 + b.ne 1b + ret +endfunc diff --git a/libavcodec/aarch64/vp9mc_neon.S b/libavcodec/aarch64/vp9mc_neon.S index f67624ca04..abf2bae9db 100644 --- a/libavcodec/aarch64/vp9mc_neon.S +++ b/libavcodec/aarch64/vp9mc_neon.S @@ -25,23 +25,6 @@ // const uint8_t *ref, ptrdiff_t ref_stride, // int h, int mx, int my); -function ff_vp9_copy64_aarch64, export=1 -1: - ldp x5, x6, [x2] - ldp x7, x8, [x2, #16] - stp x5, x6, [x0] - ldp x9, x10, [x2, #32] - stp x7, x8, [x0, #16] - subs w4, w4, #1 - ldp x11, x12, [x2, #48] - stp x9, x10, [x0, #32] - stp x11, x12, [x0, #48] - add x2, x2, x3 - add x0, x0, x1 - b.ne 1b - ret -endfunc - function ff_vp9_avg64_neon, export=1 mov x5, x0 1: @@ -64,19 +47,6 @@ function ff_vp9_avg64_neon, export=1 ret endfunc -function ff_vp9_copy32_aarch64, export=1 -1: - ldp x5, x6, [x2] - ldp x7, x8, [x2, #16] - stp x5, x6, [x0] - subs w4, w4, #1 - stp x7, x8, [x0, #16] - add x2, x2, x3 - add x0, x0, x1 - b.ne 1b - ret -endfunc - function ff_vp9_avg32_neon, export=1 1: ld1 {v2.16b, v3.16b}, [x2], x3