From patchwork Wed Dec 11 08:46:14 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: "Fu, Linjie" X-Patchwork-Id: 16724 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 921C7449900 for ; Wed, 11 Dec 2019 10:55:49 +0200 (EET) Received: from [127.0.1.1] (localhost [127.0.0.1]) by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTP id 7846B68AF24; Wed, 11 Dec 2019 10:55:49 +0200 (EET) X-Original-To: ffmpeg-devel@ffmpeg.org Delivered-To: ffmpeg-devel@ffmpeg.org Received: from mga11.intel.com (mga11.intel.com [192.55.52.93]) by ffbox0-bg.mplayerhq.hu (Postfix) with ESMTPS id DE0EE68ADB2 for ; Wed, 11 Dec 2019 10:55:42 +0200 (EET) X-Amp-Result: SKIPPED(no attachment in message) X-Amp-File-Uploaded: False Received: from fmsmga003.fm.intel.com ([10.253.24.29]) by fmsmga102.fm.intel.com with ESMTP/TLS/DHE-RSA-AES256-GCM-SHA384; 11 Dec 2019 00:55:41 -0800 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.69,301,1571727600"; d="scan'208";a="264800622" Received: from icl-dev.sh.intel.com ([10.239.158.73]) by FMSMGA003.fm.intel.com with ESMTP; 11 Dec 2019 00:55:40 -0800 From: Linjie Fu To: ffmpeg-devel@ffmpeg.org Date: Wed, 11 Dec 2019 16:46:14 +0800 Message-Id: <1576053974-1039-1-git-send-email-linjie.fu@intel.com> X-Mailer: git-send-email 2.7.4 Subject: [FFmpeg-devel] [PATCH 3/4] lavc/x86/hevc_add_res: Fix overflow in ADD_RES_SSE_8_8 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 Cc: Xu Guangxin , Linjie Fu MIME-Version: 1.0 Errors-To: ffmpeg-devel-bounces@ffmpeg.org Sender: "ffmpeg-devel" Fix overflow for coeff -32768 in function ADD_RES_SSE_8_8 with no performance drop. ./checkasm --test=hevc_add_res --bench Mainline: - hevc_add_res.add_residual [OK] hevc_add_res_8x8_8_sse2: 15.5 Add overflow test case: - hevc_add_res.add_residual [FAILED] After: - hevc_add_res.add_residual [OK] hevc_add_res_8x8_8_sse2: 15.5 Signed-off-by: Xu Guangxin Signed-off-by: Linjie Fu --- libavcodec/x86/hevc_add_res.asm | 45 ++++++++++++++++++++--------------------- 1 file changed, 22 insertions(+), 23 deletions(-) diff --git a/libavcodec/x86/hevc_add_res.asm b/libavcodec/x86/hevc_add_res.asm index 249c607..e5e9f24 100644 --- a/libavcodec/x86/hevc_add_res.asm +++ b/libavcodec/x86/hevc_add_res.asm @@ -57,32 +57,30 @@ cglobal hevc_add_residual_4_8, 3, 3, 6 RET %macro ADD_RES_SSE_8_8 0 - pxor m3, m3 - mova m4, [r1] - mova m6, [r1+16] - mova m0, [r1+32] - mova m2, [r1+48] - psubw m5, m3, m4 - psubw m7, m3, m6 - psubw m1, m3, m0 - packuswb m4, m0 - packuswb m5, m1 - psubw m3, m2 - packuswb m6, m2 - packuswb m7, m3 - movq m0, [r0] movq m1, [r0+r2] - movhps m0, [r0+r2*2] - movhps m1, [r0+r3] - paddusb m0, m4 - paddusb m1, m6 - psubusb m0, m5 - psubusb m1, m7 + punpcklbw m0, m4 + punpcklbw m1, m4 + mova m2, [r1] + mova m3, [r1+16] + paddsw m0, m2 + paddsw m1, m3 + packuswb m0, m1 + + movq m2, [r0+r2*2] + movq m3, [r0+r3] + punpcklbw m2, m4 + punpcklbw m3, m4 + mova m6, [r1+32] + mova m7, [r1+48] + paddsw m2, m6 + paddsw m3, m7 + packuswb m2, m3 + movq [r0], m0 - movq [r0+r2], m1 - movhps [r0+2*r2], m0 - movhps [r0+r3], m1 + movhps [r0+r2], m0 + movq [r0+r2*2], m2 + movhps [r0+r3], m2 %endmacro %macro ADD_RES_SSE_16_32_8 3 @@ -120,6 +118,7 @@ cglobal hevc_add_residual_4_8, 3, 3, 6 %macro TRANSFORM_ADD_8 0 ; void ff_hevc_add_residual_8_8_(uint8_t *dst, int16_t *res, ptrdiff_t stride) cglobal hevc_add_residual_8_8, 3, 4, 8 + pxor m4, m4 lea r3, [r2*3] ADD_RES_SSE_8_8 add r1, 64