From 2f31a1644587e53bc7b5381a009e611de94a65f0 Mon Sep 17 00:00:00 2001 From: James Zern Date: Wed, 8 Mar 2017 20:42:35 -0800 Subject: [PATCH] move vp9_scale_and_extend_frame_c to vp9_frame_scale.c this is similar to the x86 configuration and helps mitigate an issue with a circular dependency between this function and the ssse3 variant causing an outsized increase in binary size (~300K for chrome) chrome.dll: .text 255B000 -> 252B000 .data 7B000 -> 75000 -221184 bytes BUG=chromium:697956 Change-Id: Ic95b142ecd62dd4f1795788aa27dd8fab59b708c --- vp9/encoder/vp9_encoder.c | 38 ------------------------ vp9/encoder/vp9_frame_scale.c | 54 +++++++++++++++++++++++++++++++++++ vp9/vp9cx.mk | 1 + 3 files changed, 55 insertions(+), 38 deletions(-) create mode 100644 vp9/encoder/vp9_frame_scale.c diff --git a/vp9/encoder/vp9_encoder.c b/vp9/encoder/vp9_encoder.c index 3c926f134..03bf3850a 100644 --- a/vp9/encoder/vp9_encoder.c +++ b/vp9/encoder/vp9_encoder.c @@ -2408,44 +2408,6 @@ static void scale_and_extend_frame(const YV12_BUFFER_CONFIG *src, } #endif // CONFIG_VP9_HIGHBITDEPTH -void vp9_scale_and_extend_frame_c(const YV12_BUFFER_CONFIG *src, - YV12_BUFFER_CONFIG *dst) { - const int src_w = src->y_crop_width; - const int src_h = src->y_crop_height; - const int dst_w = dst->y_crop_width; - const int dst_h = dst->y_crop_height; - const uint8_t *const srcs[3] = { src->y_buffer, src->u_buffer, - src->v_buffer }; - const int src_strides[3] = { src->y_stride, src->uv_stride, src->uv_stride }; - uint8_t *const dsts[3] = { dst->y_buffer, dst->u_buffer, dst->v_buffer }; - const int dst_strides[3] = { dst->y_stride, dst->uv_stride, dst->uv_stride }; - const InterpKernel *const kernel = vp9_filter_kernels[EIGHTTAP]; - int x, y, i; - - for (i = 0; i < MAX_MB_PLANE; ++i) { - const int factor = (i == 0 || i == 3 ? 1 : 2); - const int src_stride = src_strides[i]; - const int dst_stride = dst_strides[i]; - for (y = 0; y < dst_h; y += 16) { - const int y_q4 = y * (16 / factor) * src_h / dst_h; - for (x = 0; x < dst_w; x += 16) { - const int x_q4 = x * (16 / factor) * src_w / dst_w; - const uint8_t *src_ptr = srcs[i] + - (y / factor) * src_h / dst_h * src_stride + - (x / factor) * src_w / dst_w; - uint8_t *dst_ptr = dsts[i] + (y / factor) * dst_stride + (x / factor); - - vpx_scaled_2d(src_ptr, src_stride, dst_ptr, dst_stride, - kernel[x_q4 & 0xf], 16 * src_w / dst_w, - kernel[y_q4 & 0xf], 16 * src_h / dst_h, 16 / factor, - 16 / factor); - } - } - } - - vpx_extend_frame_borders(dst); -} - static int scale_down(VP9_COMP *cpi, int q) { RATE_CONTROL *const rc = &cpi->rc; GF_GROUP *const gf_group = &cpi->twopass.gf_group; diff --git a/vp9/encoder/vp9_frame_scale.c b/vp9/encoder/vp9_frame_scale.c new file mode 100644 index 000000000..349e7bd41 --- /dev/null +++ b/vp9/encoder/vp9_frame_scale.c @@ -0,0 +1,54 @@ +/* + * Copyright (c) 2017 The WebM project authors. All Rights Reserved. + * + * Use of this source code is governed by a BSD-style license + * that can be found in the LICENSE file in the root of the source + * tree. An additional intellectual property rights grant can be found + * in the file PATENTS. All contributing project authors may + * be found in the AUTHORS file in the root of the source tree. + */ + +#include "./vp9_rtcd.h" +#include "./vpx_dsp_rtcd.h" +#include "./vpx_scale_rtcd.h" +#include "vp9/common/vp9_blockd.h" +#include "vpx_dsp/vpx_filter.h" +#include "vpx_scale/yv12config.h" + +void vp9_scale_and_extend_frame_c(const YV12_BUFFER_CONFIG *src, + YV12_BUFFER_CONFIG *dst) { + const int src_w = src->y_crop_width; + const int src_h = src->y_crop_height; + const int dst_w = dst->y_crop_width; + const int dst_h = dst->y_crop_height; + const uint8_t *const srcs[3] = { src->y_buffer, src->u_buffer, + src->v_buffer }; + const int src_strides[3] = { src->y_stride, src->uv_stride, src->uv_stride }; + uint8_t *const dsts[3] = { dst->y_buffer, dst->u_buffer, dst->v_buffer }; + const int dst_strides[3] = { dst->y_stride, dst->uv_stride, dst->uv_stride }; + const InterpKernel *const kernel = vp9_filter_kernels[EIGHTTAP]; + int x, y, i; + + for (i = 0; i < MAX_MB_PLANE; ++i) { + const int factor = (i == 0 || i == 3 ? 1 : 2); + const int src_stride = src_strides[i]; + const int dst_stride = dst_strides[i]; + for (y = 0; y < dst_h; y += 16) { + const int y_q4 = y * (16 / factor) * src_h / dst_h; + for (x = 0; x < dst_w; x += 16) { + const int x_q4 = x * (16 / factor) * src_w / dst_w; + const uint8_t *src_ptr = srcs[i] + + (y / factor) * src_h / dst_h * src_stride + + (x / factor) * src_w / dst_w; + uint8_t *dst_ptr = dsts[i] + (y / factor) * dst_stride + (x / factor); + + vpx_scaled_2d(src_ptr, src_stride, dst_ptr, dst_stride, + kernel[x_q4 & 0xf], 16 * src_w / dst_w, + kernel[y_q4 & 0xf], 16 * src_h / dst_h, 16 / factor, + 16 / factor); + } + } + } + + vpx_extend_frame_borders(dst); +} diff --git a/vp9/vp9cx.mk b/vp9/vp9cx.mk index fe2baf5e8..e0913bea3 100644 --- a/vp9/vp9cx.mk +++ b/vp9/vp9cx.mk @@ -39,6 +39,7 @@ VP9_CX_SRCS-yes += encoder/vp9_encodemb.h VP9_CX_SRCS-yes += encoder/vp9_encodemv.h VP9_CX_SRCS-yes += encoder/vp9_extend.h VP9_CX_SRCS-yes += encoder/vp9_firstpass.h +VP9_CX_SRCS-yes += encoder/vp9_frame_scale.c VP9_CX_SRCS-yes += encoder/vp9_job_queue.h VP9_CX_SRCS-yes += encoder/vp9_lookahead.c VP9_CX_SRCS-yes += encoder/vp9_lookahead.h -- 2.40.0