From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-yb1-xb29.google.com (mail-yb1-xb29.google.com [IPv6:2607:f8b0:4864:20::b29]) by sourceware.org (Postfix) with ESMTPS id F009E3857BB8 for ; Thu, 14 Jul 2022 03:05:12 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.1 sourceware.org F009E3857BB8 Received: by mail-yb1-xb29.google.com with SMTP id c131so324719ybf.9 for ; Wed, 13 Jul 2022 20:05:12 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20210112; h=x-gm-message-state:mime-version:references:in-reply-to:from:date :message-id:subject:to:cc; bh=hNSqb2yyrft2ik5pQqCFzHChqgzbA1GLKGdrtbo7fSE=; b=mH2FcuuBT30qG4/xsOs1oHy/eIa09BJZC0YCFdt8WtwYNif+IMC7aTzRFep5Ok8yWU iQLyQzxXd1OtMdH1JLPZ3QdBnbBrHq7dJZsqE6+UDWUNIh2DPKyxYG6QJqcAE/jAXKKc zXLZeIYAWaONkPL2r3FQ9lsWpwq2CYG+9003lE10CRVIgXrR9Sut7EbLMYjjQNxjmPlr cw75Lo2gjDnqLN0Z/ZgsMqGuI7cjNWgM3t53bar6hObLRHN50q1HRzEeX0dc5m5urP3U 90usr7sYGRub87eTuQA6aVbWcEGREvsbe1pSi7Ynu8itSm/NZ8V+zeE+RRTYvwrfZE/W Ko/w== X-Gm-Message-State: AJIora//1yUBRx3TD2jQLKEvT0hM4hshAcgGaf/H6QbeZVD5SyGmmmq8 NYurWC3yJWZXTpJKQl8uieuvTmNwe8xxqxlDao78zz68P6M= X-Google-Smtp-Source: AGRyM1vF5QIFzzfx2KzxMnxE6ra6j9WW8UDeG0v3naAt7FhiANn/Iei/6Oug1fRpD6Sa5miln3xIeoZR1ojqY5LPvJg= X-Received: by 2002:a25:5809:0:b0:66c:8709:3611 with SMTP id m9-20020a255809000000b0066c87093611mr6255581ybb.608.1657767912501; Wed, 13 Jul 2022 20:05:12 -0700 (PDT) MIME-Version: 1.0 References: <20220628152735.17863-1-goldstein.w.n@gmail.com> <20220629230706.1264225-1-goldstein.w.n@gmail.com> <20220629230706.1264225-2-goldstein.w.n@gmail.com> In-Reply-To: From: Sunil Pandey Date: Wed, 13 Jul 2022 20:04:36 -0700 Message-ID: Subject: Re: [PATCH v3 2/3] x86: Move and slightly improve memset_erms To: "H.J. Lu" Cc: Noah Goldstein , GNU C Library Content-Type: text/plain; charset="UTF-8" X-Spam-Status: No, score=-6.9 required=5.0 tests=BAYES_00, DKIM_SIGNED, DKIM_VALID, DKIM_VALID_AU, DKIM_VALID_EF, FREEMAIL_ENVFROM_END_DIGIT, FREEMAIL_FROM, GIT_PATCH_0, HK_RANDOM_ENVFROM, HK_RANDOM_FROM, KAM_SHORT, RCVD_IN_DNSWL_NONE, SPF_HELO_NONE, SPF_PASS, TXREP, T_SCC_BODY_TEXT_LINE autolearn=ham autolearn_force=no version=3.4.6 X-Spam-Checker-Version: SpamAssassin 3.4.6 (2021-04-09) on server2.sourceware.org X-BeenThere: libc-alpha@sourceware.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Libc-alpha mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 14 Jul 2022 03:05:14 -0000 On Wed, Jun 29, 2022 at 4:30 PM H.J. Lu via Libc-alpha wrote: > > On Wed, Jun 29, 2022 at 4:07 PM Noah Goldstein wrote: > > > > Implementation wise: > > 1. Remove the VZEROUPPER as memset_{impl}_unaligned_erms does not > > use the L(stosb) label that was previously defined. > > > > 2. Don't give the hotpath (fallthrough) to zero size. > > > > Code positioning wise: > > > > Move memset_{chk}_erms to its own file. Leaving it in between the > > memset_{impl}_unaligned both adds unnecessary complexity to the > > file and wastes space in a relatively hot cache section. > > --- > > sysdeps/x86_64/multiarch/Makefile | 1 + > > sysdeps/x86_64/multiarch/memset-erms.S | 44 +++++++++++++++++++ > > .../multiarch/memset-vec-unaligned-erms.S | 31 ------------- > > 3 files changed, 45 insertions(+), 31 deletions(-) > > create mode 100644 sysdeps/x86_64/multiarch/memset-erms.S > > > > diff --git a/sysdeps/x86_64/multiarch/Makefile b/sysdeps/x86_64/multiarch/Makefile > > index 62a4d96fb8..18cea04423 100644 > > --- a/sysdeps/x86_64/multiarch/Makefile > > +++ b/sysdeps/x86_64/multiarch/Makefile > > @@ -30,6 +30,7 @@ sysdep_routines += \ > > memset-avx2-unaligned-erms-rtm \ > > memset-avx512-no-vzeroupper \ > > memset-avx512-unaligned-erms \ > > + memset-erms \ > > memset-evex-unaligned-erms \ > > memset-sse2-unaligned-erms \ > > rawmemchr-avx2 \ > > diff --git a/sysdeps/x86_64/multiarch/memset-erms.S b/sysdeps/x86_64/multiarch/memset-erms.S > > new file mode 100644 > > index 0000000000..e83cccc731 > > --- /dev/null > > +++ b/sysdeps/x86_64/multiarch/memset-erms.S > > @@ -0,0 +1,44 @@ > > +/* memset implement with rep stosb > > + Copyright (C) 2022 Free Software Foundation, Inc. > > + This file is part of the GNU C Library. > > + > > + The GNU C Library is free software; you can redistribute it and/or > > + modify it under the terms of the GNU Lesser General Public > > + License as published by the Free Software Foundation; either > > + version 2.1 of the License, or (at your option) any later version. > > + > > + The GNU C Library is distributed in the hope that it will be useful, > > + but WITHOUT ANY WARRANTY; without even the implied warranty of > > + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU > > + Lesser General Public License for more details. > > + > > + You should have received a copy of the GNU Lesser General Public > > + License along with the GNU C Library; if not, see > > + . */ > > + > > + > > +#include > > + > > +#if defined USE_MULTIARCH && IS_IN (libc) > > + .text > > +ENTRY (__memset_chk_erms) > > + cmp %RDX_LP, %RCX_LP > > + jb HIDDEN_JUMPTARGET (__chk_fail) > > +END (__memset_chk_erms) > > + > > +/* Only used to measure performance of REP STOSB. */ > > +ENTRY (__memset_erms) > > + /* Skip zero length. */ > > + test %RDX_LP, %RDX_LP > > + jz L(stosb_return_zero) > > + mov %RDX_LP, %RCX_LP > > + movzbl %sil, %eax > > + mov %RDI_LP, %RDX_LP > > + rep stosb > > + mov %RDX_LP, %RAX_LP > > + ret > > +L(stosb_return_zero): > > + movq %rdi, %rax > > + ret > > +END (__memset_erms) > > +#endif > > diff --git a/sysdeps/x86_64/multiarch/memset-vec-unaligned-erms.S b/sysdeps/x86_64/multiarch/memset-vec-unaligned-erms.S > > index abc12d9cda..905d0fa464 100644 > > --- a/sysdeps/x86_64/multiarch/memset-vec-unaligned-erms.S > > +++ b/sysdeps/x86_64/multiarch/memset-vec-unaligned-erms.S > > @@ -156,37 +156,6 @@ L(entry_from_wmemset): > > #if defined USE_MULTIARCH && IS_IN (libc) > > END (MEMSET_SYMBOL (__memset, unaligned)) > > > > -# if VEC_SIZE == 16 > > -ENTRY (__memset_chk_erms) > > - cmp %RDX_LP, %RCX_LP > > - jb HIDDEN_JUMPTARGET (__chk_fail) > > -END (__memset_chk_erms) > > - > > -/* Only used to measure performance of REP STOSB. */ > > -ENTRY (__memset_erms) > > - /* Skip zero length. */ > > - test %RDX_LP, %RDX_LP > > - jnz L(stosb) > > - movq %rdi, %rax > > - ret > > -# else > > -/* Provide a hidden symbol to debugger. */ > > - .hidden MEMSET_SYMBOL (__memset, erms) > > -ENTRY (MEMSET_SYMBOL (__memset, erms)) > > -# endif > > -L(stosb): > > - mov %RDX_LP, %RCX_LP > > - movzbl %sil, %eax > > - mov %RDI_LP, %RDX_LP > > - rep stosb > > - mov %RDX_LP, %RAX_LP > > - VZEROUPPER_RETURN > > -# if VEC_SIZE == 16 > > -END (__memset_erms) > > -# else > > -END (MEMSET_SYMBOL (__memset, erms)) > > -# endif > > - > > # if defined SHARED && IS_IN (libc) > > ENTRY_CHK (MEMSET_CHK_SYMBOL (__memset_chk, unaligned_erms)) > > cmp %RDX_LP, %RCX_LP > > -- > > 2.34.1 > > > > LGTM. > > Thanks. > > -- > H.J. I would like to backport this patch to release branches. Any comments or objections? --Sunil