From patchwork Thu Jan 7 12:10:19 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Josh Dekker X-Patchwork-Id: 24830 Return-Path: X-Original-To: patchwork@ffaux-bg.ffmpeg.org Delivered-To: patchwork@ffaux-bg.ffmpeg.org Received: from ffbox0-bg.mplayerhq.hu (ffbox0-bg.ffmpeg.org [79.124.17.100]) by ffaux.localdomain (Postfix) with ESMTP id 293F544AF79 for ; Thu, 7 Jan 2021 14:10:37 +0200 (EET) Received: from [127.0.1.1] (localhost [127.0.0.1]) by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTP id 1655468A370; Thu, 7 Jan 2021 14:10:37 +0200 (EET) X-Original-To: ffmpeg-devel@ffmpeg.org Delivered-To: ffmpeg-devel@ffmpeg.org Received: from wout3-smtp.messagingengine.com (wout3-smtp.messagingengine.com [64.147.123.19]) by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTPS id EFD43688288 for ; Thu, 7 Jan 2021 14:10:28 +0200 (EET) Received: from compute3.internal (compute3.nyi.internal [10.202.2.43]) by mailout.west.internal (Postfix) with ESMTP id 6D546180E for ; Thu, 7 Jan 2021 07:10:27 -0500 (EST) Received: from mailfrontend1 ([10.202.2.162]) by compute3.internal (MEProxy); Thu, 07 Jan 2021 07:10:27 -0500 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=itanimul.li; h= from:to:subject:date:message-id:in-reply-to:references :mime-version:content-transfer-encoding; s=fm3; bh=hMfy+BKFNus/I fgQgEkmXZTwz8LdL/U9gmrGSu5zy3o=; b=SUNvHOGuKSBaD0TIk86cGBSMftWOl s3bInVVjfO5Ph4QtvnzSDjMmqhDpSr1ZYBugxkiSrhoW8i+tW38XYvB4hx7EvXuH wKJbEZFKCOWvO9ZLVS8bW5sqoghGiuRBCCPMnJXnG24bmot8tUDgBelZJ9ZcnPE4 b26FnPayKeWvjtxGbAdJaRhA7FxTY7XcH8Kve3qew/ieav2n/LSPBdGqsOzJhmAL QGyS0rBBkXvtDJYAuT38UqESZNAFbEpGybTs5+aTH6z87xVyt6lkO5V35cIa6tCi iFZnONEAcdYIbZIItHVfTAvGgY6cr4rqiFy+VQ1++6DUtjQZeQ7iRm+Tg== DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d= messagingengine.com; h=content-transfer-encoding:date:from :in-reply-to:message-id:mime-version:references:subject:to :x-me-proxy:x-me-proxy:x-me-sender:x-me-sender:x-sasl-enc; s= fm1; bh=hMfy+BKFNus/IfgQgEkmXZTwz8LdL/U9gmrGSu5zy3o=; b=KKmSgPEr Cl+Lt4K1xGDAzIvICZOogaKLC7ttUYGCMem2lX0ruoCa2Ef9NKFMEvrFtFOGBlKU qii0aGHpsWXaHNj/gNPmb+tF6p2P2gXRnQgm2CCKmELvSi+IlUW3gFlvoR3YPu0c 9bdKpVa3nZQT09fSGSoWbagaonOie9dLhYf0niYA3OgRznktcCaPIbSs5QFxRyOU iOPAfeAX2eHbhRHhP0jM+mzINaWuygrkQoE5wH6rXFNNNd8ljOThmYewAmKeoyyt LWeC1e2Mncsg2c5VP5nbtxE6w+5NTJdUHGA8CEJfA0IGjV4jgDYMp71zwjASQxDf YjZwY/HNqDHtYA== X-ME-Sender: X-ME-Proxy-Cause: gggruggvucftvghtrhhoucdtuddrgedujedrvdegvddgfeejucetufdoteggodetrfdotf fvucfrrhhofhhilhgvmecuhfgrshhtofgrihhlpdfqfgfvpdfurfetoffkrfgpnffqhgen uceurghilhhouhhtmecufedttdenucenucfjughrpefhvffufffkofgjfhgggfestdekre dtredttdenucfhrhhomheplfhoshhhucffvghkkhgvrhcuoehjohhshhesihhtrghnihhm uhhlrdhliheqnecuggftrfgrthhtvghrnhepgeduieeltdeltedugfegffehfefgudduhe ehgfejgeeuiedukedvtdejiefgheeunecuffhomhgrihhnpehnvghonhdrshgsnecukfhp peekkedrudeftddrgeekrddufeejnecuvehluhhsthgvrhfuihiivgeptdenucfrrghrrg hmpehmrghilhhfrhhomhepjhhoshhhsehithgrnhhimhhulhdrlhhi X-ME-Proxy: Received: from localhost.localdomain (mue-88-130-48-137.dsl.tropolys.de [88.130.48.137]) by mail.messagingengine.com (Postfix) with ESMTPA id 9600024005A for ; Thu, 7 Jan 2021 07:10:26 -0500 (EST) From: Josh Dekker To: ffmpeg-devel@ffmpeg.org Date: Thu, 7 Jan 2021 13:10:19 +0100 Message-Id: <20210107121020.86179-4-josh@itanimul.li> X-Mailer: git-send-email 2.24.3 (Apple Git-128) In-Reply-To: <20210107121020.86179-1-josh@itanimul.li> References: <20210107121020.86179-1-josh@itanimul.li> MIME-Version: 1.0 Subject: [FFmpeg-devel] [PATCH 3/4] lavc/aarch64: add HEVC sao_band NEON X-BeenThere: ffmpeg-devel@ffmpeg.org X-Mailman-Version: 2.1.20 Precedence: list List-Id: FFmpeg development discussions and patches List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Reply-To: FFmpeg development discussions and patches Errors-To: ffmpeg-devel-bounces@ffmpeg.org Sender: "ffmpeg-devel" Only works for 8x8. Signed-off-by: Josh Dekker --- libavcodec/aarch64/Makefile | 3 +- libavcodec/aarch64/hevcdsp_init.c | 7 +++ libavcodec/aarch64/hevcdsp_sao_neon.S | 87 +++++++++++++++++++++++++++ 3 files changed, 96 insertions(+), 1 deletion(-) create mode 100644 libavcodec/aarch64/hevcdsp_sao_neon.S diff --git a/libavcodec/aarch64/Makefile b/libavcodec/aarch64/Makefile index 42d80bf74c..1f54fc31f4 100644 --- a/libavcodec/aarch64/Makefile +++ b/libavcodec/aarch64/Makefile @@ -55,7 +55,8 @@ NEON-OBJS-$(CONFIG_VP8DSP) += aarch64/vp8dsp_neon.o NEON-OBJS-$(CONFIG_AAC_DECODER) += aarch64/aacpsdsp_neon.o NEON-OBJS-$(CONFIG_DCA_DECODER) += aarch64/synth_filter_neon.o NEON-OBJS-$(CONFIG_HEVC_DECODER) += aarch64/hevcdsp_add_res_neon.o \ - aarch64/hevcdsp_idct_neon.o + aarch64/hevcdsp_idct_neon.o \ + aarch64/hevcdsp_sao_neon.o NEON-OBJS-$(CONFIG_OPUS_DECODER) += aarch64/opusdsp_neon.o NEON-OBJS-$(CONFIG_VORBIS_DECODER) += aarch64/vorbisdsp_neon.o NEON-OBJS-$(CONFIG_VP9_DECODER) += aarch64/vp9itxfm_16bpp_neon.o \ diff --git a/libavcodec/aarch64/hevcdsp_init.c b/libavcodec/aarch64/hevcdsp_init.c index 2cd7ef3a6c..8f0a923ab1 100644 --- a/libavcodec/aarch64/hevcdsp_init.c +++ b/libavcodec/aarch64/hevcdsp_init.c @@ -23,6 +23,11 @@ #include "libavcodec/hevcdsp.h" #include "libavcodec/avcodec.h" +void ff_hevc_sao_band_filter_8x8_8_neon(uint8_t *_dst, uint8_t *_src, + ptrdiff_t stride_dst, ptrdiff_t stride_src, + int16_t *sao_offset_val, int sao_left_class, + int width, int height); + void ff_hevc_idct_4x4_dc_8_neon(int16_t *coeffs); void ff_hevc_idct_8x8_dc_8_neon(int16_t *coeffs); void ff_hevc_idct_16x16_dc_8_neon(int16_t *coeffs); @@ -53,6 +58,8 @@ av_cold void ff_hevc_dsp_init_aarch64(HEVCDSPContext *c, const int bit_depth) { int cpu_flags = av_get_cpu_flags(); if (have_neon(cpu_flags) && bit_depth == 8) { + c->sao_band_filter[0] = ff_hevc_sao_band_filter_8x8_8_neon; + c->add_residual[0] = ff_hevc_add_residual_4x4_8_neon; c->add_residual[1] = ff_hevc_add_residual_8x8_8_neon; c->add_residual[2] = ff_hevc_add_residual_16x16_8_neon; diff --git a/libavcodec/aarch64/hevcdsp_sao_neon.S b/libavcodec/aarch64/hevcdsp_sao_neon.S new file mode 100644 index 0000000000..25b6c25117 --- /dev/null +++ b/libavcodec/aarch64/hevcdsp_sao_neon.S @@ -0,0 +1,87 @@ +/* -*-arm64-*- + * vim: syntax=arm64asm + * + * AArch64 NEON optimised SAO functions for HEVC decoding + * + * Copyright (c) 2020 Josh Dekker + * + * 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" + +// void sao_band_filter(uint8_t *_dst, uint8_t *_src, +// ptrdiff_t stride_dst, ptrdiff_t stride_src, +// int16_t *sao_offset_val, int sao_left_class, +// int width, int height) +function ff_hevc_sao_band_filter_8x8_8_neon, export=1 + sub sp, sp, #64 + stp xzr, xzr, [sp] + stp xzr, xzr, [sp, #16] + stp xzr, xzr, [sp, #32] + stp xzr, xzr, [sp, #48] + mov w8, #4 +.setup: + ldrsh x9, [x4, x8, lsl #1] // x9 = sao_offset_val[k+1] + subs w8, w8, #1 + add w10, w8, w5 // x10 = k + sao_left_class + and w10, w10, #0x1F + strh w9, [sp, x10, lsl #1] + bne .setup + ld1 {v16.16B-v19.16B}, [sp], #64 + movi v20.8H, #1 +0: // beginning of line + mov w8, w6 +8: + // Simple layout for accessing 16bit values + // with 8bit LUT. + // + // 00 01 02 03 04 05 06 07 + // +-----------------------------------> + // |xDE#xAD|xCA#xFE|xBE#xEF|xFE#xED|.... + // +-----------------------------------> + // i-0 i-1 i-2 i-3 + // dst[x] = av_clip_pixel(src[x] + offset_table[src[x] >> shift]); + ld1 {v2.8B}, [x1] + // load src[x] + ushll v0.8H, v2.8B, #0 + // >> shift + ushr v2.8H, v0.8H, #3 // BIT_DEPTH - 3 + // x2 (access lower short) + shl v1.8H, v2.8H, #1 // low (x2, accessing short) + // +1 access upper short + add v3.8H, v1.8H, v20.8H + // shift insert index to upper byte + sli v1.8H, v3.8H, #8 + // table + tbx v2.16B, {v16.16B-v19.16B}, v1.16B + // src[x] + table + add v1.8H, v0.8H, v2.8H + // clip + narrow + sqxtun v4.8B, v1.8H + // store + st1 {v4.8B}, [x0] + // done 8 pixels + subs w8, w8, #8 + bne 8b + // finished line + subs w7, w7, #1 + add x0, x0, x2 // dst += stride_dst + add x1, x1, x3 // src += stride_src + bne 0b + ret +endfunc