From patchwork Tue Jan 10 21:01:02 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Adhemerval Zanella Netto X-Patchwork-Id: 640901 Delivered-To: patch@linaro.org Received: by 2002:a17:522:f3c4:b0:4b4:3859:abed with SMTP id in4csp2916652pvb; Tue, 10 Jan 2023 13:02:55 -0800 (PST) X-Google-Smtp-Source: AMrXdXvDJPdnBDe/+evkPXne4pUfItMXxZeThsjV5E02Mqr2FeOyJTuoxU7opyKYELjyX3lHAZ+h X-Received: by 2002:a17:907:9d0b:b0:78d:f455:30db with SMTP id kt11-20020a1709079d0b00b0078df45530dbmr62547648ejc.3.1673384575384; Tue, 10 Jan 2023 13:02:55 -0800 (PST) ARC-Seal: i=1; a=rsa-sha256; t=1673384575; cv=none; d=google.com; s=arc-20160816; b=bHL9wi0f25IaL59+ec0s6yc0e/EUa3quJLNHhJgrEXQTY/A3Q91pvcsL1R5yRzaEml IHvIPSN0PQC6e/o9KcoL0/cCEbaKnoV1YxL163RS7w/WQ4Dr5avGXopfdfH1+Bfme7hl Fv/C8Luhpt3IVc07n3dQtLAorn6WzkbsAQJn0vXys3abQfV4rdqyf/JLLxaPSAH+hM+D pwa9IMtQkf6ThzVcKF0kpOLUgmK5zzDY3u5Zqk2zYMidDxFhazPvhzr2bkDYbIWstJTx ad1PFMzkBuDw4SK6dJoCM90nemtURLHg6HvBdvGlGsrj7MR876JN+vl7lxto4S/GbGzu YD8g== ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=arc-20160816; h=sender:errors-to:reply-to:from:list-subscribe:list-help:list-post :list-archive:list-unsubscribe:list-id:precedence :content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:to:dmarc-filter:delivered-to:dkim-signature :dkim-filter; bh=hWOcUJjSqHCMrNrz/r9938jmbfA0HbRSV45xxOnIXTs=; b=k8ClL0BOtVFkWRF8twnXvRGupTsT3FPclT4shWFb3GwGpCDoZo+HpFb9GcTH069/Wt b3AqEajeD8Jb8XDxLVLAXdon9DJrnuA0eNovf1sGBK9HY6Hvrljqr5oqm/O2PmbuppfG QDJiNUMZYAhW3nbavoBaFkCWySRVJfzt/gDUUa3tSYZoK2SqU4/WsQed+RBtJROidhZ8 3hisVN2ubhuyY7N2YsiIzbwew1BZ8kvFPq4UC0HfRPwKevxXM9ARa9j56vXuvf0c/kBx jwixV87SCdSHd+tqX3rp9VqhKU8xyJZgXTKFxMIk6a45tg1OSY+D4vLCCNDprPgmMpBZ x09A== ARC-Authentication-Results: i=1; mx.google.com; dkim=pass header.i=@sourceware.org header.s=default header.b=rcGtg4qu; spf=pass (google.com: domain of libc-alpha-bounces+patch=linaro.org@sourceware.org designates 8.43.85.97 as permitted sender) smtp.mailfrom="libc-alpha-bounces+patch=linaro.org@sourceware.org"; dmarc=pass (p=NONE sp=NONE dis=NONE) header.from=sourceware.org Return-Path: Received: from sourceware.org (server2.sourceware.org. [8.43.85.97]) by mx.google.com with ESMTPS id l5-20020a170906794500b007adcc8fb7a0si14755140ejo.399.2023.01.10.13.02.55 for (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Tue, 10 Jan 2023 13:02:55 -0800 (PST) Received-SPF: pass (google.com: domain of libc-alpha-bounces+patch=linaro.org@sourceware.org designates 8.43.85.97 as permitted sender) client-ip=8.43.85.97; Authentication-Results: mx.google.com; dkim=pass header.i=@sourceware.org header.s=default header.b=rcGtg4qu; spf=pass (google.com: domain of libc-alpha-bounces+patch=linaro.org@sourceware.org designates 8.43.85.97 as permitted sender) smtp.mailfrom="libc-alpha-bounces+patch=linaro.org@sourceware.org"; dmarc=pass (p=NONE sp=NONE dis=NONE) header.from=sourceware.org Received: from server2.sourceware.org (localhost [IPv6:::1]) by sourceware.org (Postfix) with ESMTP id 3FA6438432E6 for ; Tue, 10 Jan 2023 21:02:54 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 3FA6438432E6 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=sourceware.org; s=default; t=1673384574; bh=hWOcUJjSqHCMrNrz/r9938jmbfA0HbRSV45xxOnIXTs=; h=To:Subject:Date:In-Reply-To:References:List-Id:List-Unsubscribe: List-Archive:List-Post:List-Help:List-Subscribe:From:Reply-To: From; b=rcGtg4quF8Ic+jAD+U9kGAMMrw9zuHpmmm0RsA4xgiqsr3MByQaq1/tuNfXv9p9Mj 6XeEOKe2kWZC3PvI5MPC+XsIQbgFK8XOFy9/tjTDgaWet4yfTFmp3Adq8H6khlbeLL Zwvwwdg4FocZar3oRkW+uIOMAWo6xqmvOhWRuHKs= X-Original-To: libc-alpha@sourceware.org Delivered-To: libc-alpha@sourceware.org Received: from mail-oo1-xc36.google.com (mail-oo1-xc36.google.com [IPv6:2607:f8b0:4864:20::c36]) by sourceware.org (Postfix) with ESMTPS id F100838493E9 for ; Tue, 10 Jan 2023 21:01:44 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.2 sourceware.org F100838493E9 Received: by mail-oo1-xc36.google.com with SMTP id c190-20020a4a4fc7000000b004a3addd10b5so3551045oob.1 for ; Tue, 10 Jan 2023 13:01:44 -0800 (PST) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20210112; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:to:from:x-gm-message-state:from:to:cc :subject:date:message-id:reply-to; bh=hWOcUJjSqHCMrNrz/r9938jmbfA0HbRSV45xxOnIXTs=; b=FhpqT66gVaHRSfpb2Nnk+yyw+eJZrjwCJ0+WS8OJBUHXt8y+MJstm/oZ2sDVsK8Ka5 DN+eJCaE2X29Zqr8o8Tq6rO4Z5xf610w3ZrS+58e9bBScP2ScpK0MyPH/DdWAQ62O689 5UeNXiMCh4WOStqajxQMOZ2FLK+ZUbvTU1EEsOQfZK3BZmNYQBtCRWz+f+X+RJHAAKSG KiqSNQcDH/MwPJKf3PNTJlnDsIAcyxXdWqd0nk93eieeHiFIcEJtlyINhUadXPzv/Aot YgP55OgcC6GJIqCNzse8BnEkHiPOHQZd0op1RjBHr13CqMCN+7Z8rz1VfjfoMw/BMI50 zw5Q== X-Gm-Message-State: AFqh2kp1SHS9ZdQpxTMvQ1zhHTmax7r2DLM4MiXAB9UhRY//5adfApIM C5sElnqFDxCghQYt3I1EsNkf1T7p6wSsm2XkD0Q= X-Received: by 2002:a4a:4b07:0:b0:4a0:23af:db58 with SMTP id q7-20020a4a4b07000000b004a023afdb58mr29717140ooa.7.1673384503510; Tue, 10 Jan 2023 13:01:43 -0800 (PST) Received: from mandiga.. ([2804:1b3:a7c0:a93a:e8a0:dd55:3328:997]) by smtp.gmail.com with ESMTPSA id r5-20020a4a83c5000000b0049ee88e86f9sm6202193oog.10.2023.01.10.13.01.41 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Tue, 10 Jan 2023 13:01:42 -0800 (PST) To: libc-alpha@sourceware.org, Noah Goldstein , Richard Henderson Subject: [PATCH v6 13/17] hppa: Add string-fzb.h and string-fzi.h Date: Tue, 10 Jan 2023 18:01:02 -0300 Message-Id: <20230110210106.1457686-14-adhemerval.zanella@linaro.org> X-Mailer: git-send-email 2.34.1 In-Reply-To: <20230110210106.1457686-1-adhemerval.zanella@linaro.org> References: <20230110210106.1457686-1-adhemerval.zanella@linaro.org> MIME-Version: 1.0 X-Spam-Status: No, score=-12.8 required=5.0 tests=BAYES_00, DKIM_SIGNED, DKIM_VALID, DKIM_VALID_AU, DKIM_VALID_EF, GIT_PATCH_0, KAM_SHORT, RCVD_IN_DNSWL_NONE, SPF_HELO_NONE, SPF_PASS, TXREP 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-Patchwork-Original-From: Adhemerval Zanella via Libc-alpha From: Adhemerval Zanella Netto Reply-To: Adhemerval Zanella Errors-To: libc-alpha-bounces+patch=linaro.org@sourceware.org Sender: "Libc-alpha" From: Richard Henderson Use UXOR,SBZ to test for a zero byte within a word. While we can get semi-decent code out of asm-goto, we would do slightly better with a compiler builtin. For index_zero et al, sequential testing of bytes is less expensive than any tricks that involve a count-leading-zeros insn that we don't have. Checked on hppa-linux-gnu. --- sysdeps/hppa/string-fzb.h | 70 +++++++++++++++++++ sysdeps/hppa/string-fzi.h | 139 ++++++++++++++++++++++++++++++++++++++ 2 files changed, 209 insertions(+) create mode 100644 sysdeps/hppa/string-fzb.h create mode 100644 sysdeps/hppa/string-fzi.h diff --git a/sysdeps/hppa/string-fzb.h b/sysdeps/hppa/string-fzb.h new file mode 100644 index 0000000000..865e548492 --- /dev/null +++ b/sysdeps/hppa/string-fzb.h @@ -0,0 +1,70 @@ +/* Zero byte detection, boolean. HPPA version. + Copyright (C) 2023 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 + . */ + +#ifndef _STRING_FZB_H +#define _STRING_FZB_H 1 + +#include +#include + +/* Determine if any byte within X is zero. This is a pure boolean test. */ + +static __always_inline _Bool +has_zero (op_t x) +{ + _Static_assert (sizeof (op_t) == 4, "64-bit not supported"); + + /* It's more useful to expose a control transfer to the compiler + than to expose a proper boolean result. */ + asm goto ("uxor,sbz %%r0,%0,%%r0\n\t" + "b,n %l1" : : "r"(x) : : nbz); + return 1; + nbz: + return 0; +} + +/* Likewise, but for byte equality between X1 and X2. */ + +static __always_inline _Bool +has_eq (op_t x1, op_t x2) +{ + _Static_assert (sizeof (op_t) == 4, "64-bit not supported"); + + asm goto ("uxor,sbz %0,%1,%%r0\n\t" + "b,n %l2" : : "r"(x1), "r"(x2) : : nbz); + return 1; + nbz: + return 0; +} + +/* Likewise, but for zeros in X1 and equal bytes between X1 and X2. */ + +static __always_inline _Bool +has_zero_eq (op_t x1, op_t x2) +{ + _Static_assert (sizeof (op_t) == 4, "64-bit not supported"); + + asm goto ("uxor,sbz %%r0,%0,%%r0\n\t" + "uxor,nbz %0,%1,%%r0\n\t" + "b,n %l2" : : "r"(x1), "r"(x2) : : sbz); + return 0; + sbz: + return 1; +} + +#endif /* _STRING_FZB_H */ diff --git a/sysdeps/hppa/string-fzi.h b/sysdeps/hppa/string-fzi.h new file mode 100644 index 0000000000..1824cbba5a --- /dev/null +++ b/sysdeps/hppa/string-fzi.h @@ -0,0 +1,139 @@ +/* string-fzi.h -- zero byte detection; indexes. HPPA version. + 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 + . */ + +#ifndef _STRING_FZI_H +#define _STRING_FZI_H 1 + +#include +#include + +_Static_assert (sizeof (op_t) == 4, "64-bit not supported"); + +static __always_inline unsigned int +index_first (op_t c) +{ + /* Since we have no clz insn, direct tests of the bytes is faster + than loading up the constants to do the masking. */ + if (c & 0xff000000) + return 0; + if (c & 0x00ff0000) + return 1; + if (c & 0x0000ff00) + return 2; + return 3; +} + +/* Given a word X that is known to contain a zero byte, return the + index of the first such within the long in memory order. */ +static __always_inline unsigned int +index_first_zero (op_t x) +{ + unsigned int ret; + + /* Since we have no clz insn, direct tests of the bytes is faster + than loading up the constants to do the masking. */ + asm ("extrw,u,<> %1,23,8,%%r0\n\t" + "ldi 2,%0\n\t" + "extrw,u,<> %1,15,8,%%r0\n\t" + "ldi 1,%0\n\t" + "extrw,u,<> %1,7,8,%%r0\n\t" + "ldi 0,%0" + : "=r"(ret) : "r"(x), "0"(3)); + + return ret; +} + +/* Similarly, but perform the search for byte equality between X1 and X2. */ +static __always_inline unsigned int +index_first_eq (op_t x1, op_t x2) +{ + return index_first_zero (x1 ^ x2); +} + +/* Similarly, but perform the search for zero within X1 or + equality between X1 and X2. */ +static __always_inline unsigned int +index_first_zero_eq (op_t x1, op_t x2) +{ + unsigned int ret; + + /* Since we have no clz insn, direct tests of the bytes is faster + than loading up the constants to do the masking. */ + asm ("extrw,u,= %1,23,8,%%r0\n\t" + "extrw,u,<> %2,23,8,%%r0\n\t" + "ldi 2,%0\n\t" + "extrw,u,= %1,15,8,%%r0\n\t" + "extrw,u,<> %2,15,8,%%r0\n\t" + "ldi 1,%0\n\t" + "extrw,u,= %1,7,8,%%r0\n\t" + "extrw,u,<> %2,7,8,%%r0\n\t" + "ldi 0,%0" + : "=r"(ret) : "r"(x1), "r"(x1 ^ x2), "0"(3)); + + return ret; +} + +/* Similarly, but perform the search for zero within X1 or + inequality between X1 and X2. */ +static __always_inline unsigned int +index_first_zero_ne (op_t x1, op_t x2) +{ + unsigned int ret; + + /* Since we have no clz insn, direct tests of the bytes is faster + than loading up the constants to do the masking. */ + asm ("extrw,u,<> %2,23,8,%%r0\n\t" + "extrw,u,<> %1,23,8,%%r0\n\t" + "ldi 2,%0\n\t" + "extrw,u,<> %2,15,8,%%r0\n\t" + "extrw,u,<> %1,15,8,%%r0\n\t" + "ldi 1,%0\n\t" + "extrw,u,<> %2,7,8,%%r0\n\t" + "extrw,u,<> %1,7,8,%%r0\n\t" + "ldi 0,%0" + : "=r"(ret) : "r"(x1), "r"(x1 ^ x2), "0"(3)); + + return ret; +} + +/* Similarly, but search for the last zero within X. */ +static __always_inline unsigned int +index_last_zero (op_t x) +{ + unsigned int ret; + + /* Since we have no ctz insn, direct tests of the bytes is faster + than loading up the constants to do the masking. */ + asm ("extrw,u,<> %1,15,8,%%r0\n\t" + "ldi 1,%0\n\t" + "extrw,u,<> %1,23,8,%%r0\n\t" + "ldi 2,%0\n\t" + "extrw,u,<> %1,31,8,%%r0\n\t" + "ldi 3,%0" + : "=r"(ret) : "r"(x), "0"(0)); + + return ret; +} + +static __always_inline unsigned int +index_last_eq (op_t x1, op_t x2) +{ + return index_last_zero (x1 ^ x2); +} + +#endif /* _STRING_FZI_H */