From patchwork Tue Jan 10 21:01:01 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: 62912 Return-Path: X-Original-To: patchwork@sourceware.org Delivered-To: patchwork@sourceware.org Received: from server2.sourceware.org (localhost [IPv6:::1]) by sourceware.org (Postfix) with ESMTP id 3AEA4388B6AE for ; Tue, 10 Jan 2023 21:03:10 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 3AEA4388B6AE DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=sourceware.org; s=default; t=1673384590; bh=uR+tmviJAnIQt3Uy9C/8G52NzH6k6qWZDh0/RH+zyro=; 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=DieTOQ6XaF/+XvaMXQPVqhSk4Gjfwl0d+knoLkleKEUcwVUTWmuVxk3RjX4ErMPBL kSTlmaKfukEYs5uKcuhY+h8KHwL9xmePfy8dKSywqWKkW3zeO6e+9vCgG4iTHnG5Ue 3lMKOXOtYXAj9FjFhtduLjomowk9glGZvS1VW+i4= X-Original-To: libc-alpha@sourceware.org Delivered-To: libc-alpha@sourceware.org Received: from mail-oo1-xc29.google.com (mail-oo1-xc29.google.com [IPv6:2607:f8b0:4864:20::c29]) by sourceware.org (Postfix) with ESMTPS id D49D13858281 for ; Tue, 10 Jan 2023 21:01:43 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.2 sourceware.org D49D13858281 Received: by mail-oo1-xc29.google.com with SMTP id q20-20020a4a3314000000b004f2177e6b38so601279ooq.3 for ; Tue, 10 Jan 2023 13:01:43 -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=uR+tmviJAnIQt3Uy9C/8G52NzH6k6qWZDh0/RH+zyro=; b=r47wAbb4Zb0DhJKaSYrA88SYbfspY7wru8yg4/bnClOx7Kv+IsRy9avcz0wAajV43j ET9j9G7SI1GJlt7ogWOnoSfrhRW5aldjpi6zlkE84fsZi9cdfnE4+DuoR6ik5EDRjKhP U7qE99oZpU31qNMRRMCjlW98MJpK4FkD1u4EoQuU4f8Tbq934FiUO8Juev29yHvURPxg c6cI4fuUELAs0QQeD6C4r6NzN7uXOqmdXGPnPnbHCs99quA/JlXXtk4L0qEX6/JKjyln l+iM3P+rxUYIebppymhbIdaU7R+5mzu45Q4JT4nuybWj65uELCCw9j1jfjawFgPvzcmC BOug== X-Gm-Message-State: AFqh2kp9Cg/DvIy2X7GBTNynNWL8KPNZn+ppu+sDfmOnR9XLPHjpX/bf IGfaR5HPU8hWqvZRuAKlZvhPKXgPy8b3SwQLxTQ= X-Google-Smtp-Source: AMrXdXsQdX8TPhk5EoCAzlSSC9wsuupKcJfpAOR3Ag/b5y3bLTB407DcRZhiPQ9CIX7fvAZCVzotfw== X-Received: by 2002:a4a:952f:0:b0:49f:8941:ffed with SMTP id m44-20020a4a952f000000b0049f8941ffedmr30671136ooi.9.1673384501451; Tue, 10 Jan 2023 13:01:41 -0800 (PST) Received: from mandiga.. ([2804:1b3:a7c0:a93a:e8a0:dd55:3328:997]) by smtp.gmail.com with ESMTPSA id r5-20020a4a83c5000000b0049ee88e86f9sm6202193oog.10.2023.01.10.13.01.39 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Tue, 10 Jan 2023 13:01:40 -0800 (PST) To: libc-alpha@sourceware.org, Noah Goldstein , Richard Henderson Subject: [PATCH v6 12/17] hppa: Add memcopy.h Date: Tue, 10 Jan 2023 18:01:01 -0300 Message-Id: <20230110210106.1457686-13-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=-12.7 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+patchwork=sourceware.org@sourceware.org Sender: "Libc-alpha" From: Richard Henderson GCC's combine pass cannot merge (x >> c | y << (32 - c)) into a double-word shift unless (1) the subtract is in the same basic block and (2) the result of the subtract is used exactly once. Neither condition is true for any use of MERGE. By forcing the use of a double-word shift, we not only reduce contention on SAR, but also allow the setting of SAR to be hoisted outside of a loop. Checked on hppa-linux-gnu. --- sysdeps/hppa/memcopy.h | 42 ++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 42 insertions(+) create mode 100644 sysdeps/hppa/memcopy.h diff --git a/sysdeps/hppa/memcopy.h b/sysdeps/hppa/memcopy.h new file mode 100644 index 0000000000..0d4b4ac435 --- /dev/null +++ b/sysdeps/hppa/memcopy.h @@ -0,0 +1,42 @@ +/* Definitions for memory copy functions, PA-RISC 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 + . */ + +#include + +/* Use a single double-word shift instead of two shifts and an ior. + If the uses of MERGE were close to the computation of shl/shr, + the compiler might have been able to create this itself. + But instead that computation is well separated. + + Using an inline function instead of a macro is the easiest way + to ensure that the types are correct. */ + +#undef MERGE + +static __always_inline op_t +MERGE (op_t w0, int shl, op_t w1, int shr) +{ + _Static_assert (OPSIZ == 4 || OPSIZ == 8, "Invalid OPSIZE"); + + op_t res; + if (OPSIZ == 4) + asm ("shrpw %1,%2,%%sar,%0" : "=r"(res) : "r"(w0), "r"(w1), "q"(shr)); + else if (OPSIZ == 8) + asm ("shrpd %1,%2,%%sar,%0" : "=r"(res) : "r"(w0), "r"(w1), "q"(shr)); + return res; +}