From nobody Tue Jul 28 17:56:27 2026 Delivered-To: importer@patchew.org Authentication-Results: mx.zohomail.com; dkim=pass; spf=pass (zohomail.com: domain of gnu.org designates 209.51.188.17 as permitted sender) smtp.mailfrom=qemu-devel-bounces+importer=patchew.org@nongnu.org; dmarc=pass(p=none dis=none) header.from=linaro.org ARC-Seal: i=1; a=rsa-sha256; t=1778977672; cv=none; d=zohomail.com; s=zohoarc; b=n+c2OxEjykumTIfviUFnUudnlBF7a3+OLPmZJ8so3TLmYG0ZIDkL1KR3A/75BRE+IzecnJ27+2BGCacgBOe8XmNPx4LknqO50MaxCIBzLTw89ESn6M3tN44VBPQAFowurlGhDolprwx4rloEyeR7E5aCWmu4faIx13wMCASG0k8= ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=zohomail.com; s=zohoarc; t=1778977672; h=Content-Transfer-Encoding:Date:Date:From:From:In-Reply-To:List-Subscribe:List-Post:List-Id:List-Archive:List-Help:List-Unsubscribe:MIME-Version:Message-ID:References:Sender:Subject:Subject:To:To:Message-Id:Reply-To:Cc; bh=x9rWSaQvjbDXBnfVxCErc01A7AXki6hbY0KLzGZpXR4=; b=GBtdITGMR32O2iNp6jXhhStSTFrOurWkC2gkv+Z+MYfaBKdwq4N9ilNoz4vz3MFcJZISWSRXHB061vkRliB2bC/uGauyBEDjRLezTArrBCdwD7NOxeZzZ6nYurEsUG59B1k/D30RZ2MrBjf7OwHtZp0id0W+ahzxHHcziIvcCak= ARC-Authentication-Results: i=1; mx.zohomail.com; dkim=pass; spf=pass (zohomail.com: domain of gnu.org designates 209.51.188.17 as permitted sender) smtp.mailfrom=qemu-devel-bounces+importer=patchew.org@nongnu.org; dmarc=pass header.from= (p=none dis=none) Return-Path: Received: from lists1p.gnu.org (lists1p.gnu.org [209.51.188.17]) by mx.zohomail.com with SMTPS id 177897767215154.13925428846687; Sat, 16 May 2026 17:27:52 -0700 (PDT) Received: from localhost ([::1] helo=lists1p.gnu.org) by lists1p.gnu.org with esmtp (Exim 4.90_1) (envelope-from ) id 1wOPKp-0001iF-Pk; Sat, 16 May 2026 20:26:23 -0400 Received: from eggs.gnu.org ([2001:470:142:3::10]) by lists1p.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.90_1) (envelope-from ) id 1wOPKR-0001dE-TW for qemu-devel@nongnu.org; Sat, 16 May 2026 20:26:01 -0400 Received: from mail-pf1-x429.google.com ([2607:f8b0:4864:20::429]) by eggs.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_128_GCM_SHA256:128) (Exim 4.90_1) (envelope-from ) id 1wOPKP-0002QL-NB for qemu-devel@nongnu.org; Sat, 16 May 2026 20:25:59 -0400 Received: by mail-pf1-x429.google.com with SMTP id d2e1a72fcca58-839dc688d6cso394308b3a.2 for ; Sat, 16 May 2026 17:25:56 -0700 (PDT) Received: from stoup.. ([71.212.142.33]) by smtp.gmail.com with ESMTPSA id d2e1a72fcca58-83f19c5be71sm9670317b3a.39.2026.05.16.17.25.54 for (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Sat, 16 May 2026 17:25:54 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=linaro.org; s=google; t=1778977555; x=1779582355; darn=nongnu.org; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:to:from:from:to:cc:subject:date:message-id :reply-to; bh=x9rWSaQvjbDXBnfVxCErc01A7AXki6hbY0KLzGZpXR4=; b=dFuYeKOj3ej/Ecm582uQwtEFQKMsI+Qyj7WmTZ79vkaqIOBjPjn2nDtxwMuBrlyRtv 5sQoGfpo4ScQyFlkWWqBb0bQf4SEySEOL6aygi1+GxJFl8m5vfg9GWPPBO8WwUke1sGv x11VZcSESZVgAqLWmMrermXGSwNBTK0PSCDwTz7k8XC2XSdkUOIpUwnvIDZgyqVVa4f8 lAO82gdaCUXACSG6ReCZtp0EZuWx5eQmpfTM4OQEjNoknNBC7RJMHrtfa3KjkLk08CCn 1taZekhmKNk2kRKQFycjHj011hNHsQ0ibcaBDqHTZhr/KN4xDPmGc1Dd7ujpnapB3VX5 R53g== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20251104; t=1778977555; x=1779582355; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:to:from:x-gm-gg:x-gm-message-state:from:to :cc:subject:date:message-id:reply-to; bh=x9rWSaQvjbDXBnfVxCErc01A7AXki6hbY0KLzGZpXR4=; b=RIaH+FIZnsszavRH+fLRbQFtRG+VXvAva+cBH2kQMoWZKiyqV3UqA1jC1Fr0a1Wi1R h7i3P1oCvRouKNJN1fvZJHjMLUw4HrT6qovfKlLOcd+OZyc+VZlG2DT1dnKRt3GojnWr GYuErkFvT3O46zUFxmPSCtJoCHfqGTP2xnkdJR/hEiFyKgF2VEGa77obdBI7yfHBytXp PoKN961mVQtcP08bzickApCMYIaDMCE9rY62+wNS+JY2hjADzS7AuMa+wRyAbrDqHen0 UQCKMjLPJu4/Ss/CLZUIDZH44Ob2xhQ3vujMA+U3mbDU5c/Jg+4zrulhsN4M0kephAoq 8InA== X-Gm-Message-State: AOJu0YxON/70rfUML4rvdON4arecAUPwjp2xHzksoBsTbRoAsXwuDsKa d3ifqKLnqU+ndDQSTVKNGru0+/owb22IyVOxEumqAk0pSaMqeXlpUzvzh/YdHcvPGmLp88yhDVW WK/Ev X-Gm-Gg: Acq92OHX1x9WItJ36iREY3iHBTmFW3LOGECY7UZ4/L7vED9F9HN5rUefmPELdhZfX68 Q8/j1HATOT+1eLTrgvXTSXodugn/YC1vNsDWY5i0mKYU58k0wOYvrjgHQRFozH2UCuA1aHGBm8C udYoOQ4pCtX5Fq5XZTygGzLaI6X7hwG+WFe3k/6hpnCyon3vDy3OnrX/f6KJEfmkzI17Z0W4BD+ b7jmhGkiNMhQgx2ulEq/AY+/mI2fKEFzbriBh9FdaAcCcu4sNzYwFRItYKIvM8WmopxEbIIQ1Np hyUnTC2XX+WhyGt4+MjeCgI0ncsaCjEm+NE3stz1FrG7GJtpGi0RzqyMFbk6Gpg+L4o4+bqhGQ8 HJvoaWLs4I92LO8nBlQRH4Gpnr4HceczEL0zV6pHq3nQ1+/15Mfh+Ud9GOwVgqfo//Oj1d+GprY qbioDac82UPX5F9K2zjfM/182zoOFa X-Received: by 2002:a05:6a00:1908:b0:834:df57:9d67 with SMTP id d2e1a72fcca58-83f33cf0bddmr10153972b3a.32.1778977554992; Sat, 16 May 2026 17:25:54 -0700 (PDT) From: Richard Henderson To: qemu-devel@nongnu.org Subject: [PATCH v4 02/12] fpu: Reorganize partsN(muladd) Date: Sat, 16 May 2026 17:25:39 -0700 Message-ID: <20260517002550.321291-3-richard.henderson@linaro.org> X-Mailer: git-send-email 2.43.0 In-Reply-To: <20260517002550.321291-1-richard.henderson@linaro.org> References: <20260517002550.321291-1-richard.henderson@linaro.org> MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Received-SPF: pass (zohomail.com: domain of gnu.org designates 209.51.188.17 as permitted sender) client-ip=209.51.188.17; envelope-from=qemu-devel-bounces+importer=patchew.org@nongnu.org; helo=lists1p.gnu.org; Received-SPF: pass client-ip=2607:f8b0:4864:20::429; envelope-from=richard.henderson@linaro.org; helo=mail-pf1-x429.google.com X-Spam_score_int: -20 X-Spam_score: -2.1 X-Spam_bar: -- X-Spam_report: (-2.1 / 5.0 requ) BAYES_00=-1.9, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, DKIM_VALID_EF=-0.1, RCVD_IN_DNSWL_NONE=-0.0001, SPF_HELO_NONE=0.001, SPF_PASS=-0.001 autolearn=ham autolearn_force=no X-Spam_action: no action X-BeenThere: qemu-devel@nongnu.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: qemu development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: qemu-devel-bounces+importer=patchew.org@nongnu.org Sender: qemu-devel-bounces+importer=patchew.org@nongnu.org X-ZohoMail-DKIM: pass (identity @linaro.org) X-ZM-MESSAGEID: 1778977672930158500 Content-Type: text/plain; charset="utf-8" Check the likely case of normal product and normal or zero addend first; shift NaN and infinity detection down; end with zero product + addend. Signed-off-by: Richard Henderson --- fpu/softfloat-parts.c.inc | 156 +++++++++++++++++--------------------- 1 file changed, 71 insertions(+), 85 deletions(-) diff --git a/fpu/softfloat-parts.c.inc b/fpu/softfloat-parts.c.inc index 3a9c2748cd..ca2d7a15c9 100644 --- a/fpu/softfloat-parts.c.inc +++ b/fpu/softfloat-parts.c.inc @@ -681,11 +681,47 @@ static FloatPartsN *partsN(muladd)(FloatPartsN *a, Fl= oatPartsN *b, FloatPartsN *c, int flags, float_status *s) { - int ab_mask, abc_mask; - FloatPartsW p_widen, c_widen; + int ab_mask =3D float_cmask(a->cls) | float_cmask(b->cls); + int c_mask =3D float_cmask(c->cls); + int abc_mask =3D ab_mask | c_mask; + bool c_sign =3D c->sign ^ !!(flags & float_muladd_negate_c); + bool p_sign =3D a->sign ^ b->sign ^ !!(flags & float_muladd_negate_pro= duct); =20 - ab_mask =3D float_cmask(a->cls) | float_cmask(b->cls); - abc_mask =3D float_cmask(c->cls) | ab_mask; + /* + * The "likely" case is A and B normal, so that the product is normal, + * and C normal or zero so that the result is normal. + */ + int likely_mask =3D ab_mask | (c_mask & ~float_cmask_zero); + if (likely(cmask_is_only_normals(likely_mask))) { + record_denormals_used(abc_mask, s); + + /* Perform the multiplication step. */ + FloatPartsW p_widen =3D { .sign =3D p_sign, .exp =3D a->exp + b->e= xp + 1 }; + fracN(mulw)(&p_widen, a, b); + if (!(p_widen.frac_hi & DECOMPOSED_IMPLICIT_BIT)) { + fracW(add)(&p_widen, &p_widen, &p_widen); + p_widen.exp -=3D 1; + } + + /* Perform the addition step. */ + if (!(c_mask & float_cmask_zero)) { + /* Zero-extend C to less significant bits. */ + FloatPartsW c_widen =3D { .sign =3D c_sign, .exp =3D c->exp }; + fracN(widen)(&c_widen, c); + + if (p_sign =3D=3D c_sign) { + partsW(add_normal)(&p_widen, &c_widen); + } else if (!partsW(sub_normal)(&p_widen, &c_widen)) { + goto return_sub_zero; + } + } + + /* Narrow with sticky bit, for proper rounding later. */ + fracN(truncjam)(a, &p_widen); + a->sign =3D p_widen.sign; + a->exp =3D p_widen.exp; + return a; + } =20 /* * It is implementation-defined whether the cases of (0,inf,qnan) @@ -698,97 +734,47 @@ static FloatPartsN *partsN(muladd)(FloatPartsN *a, Fl= oatPartsN *b, return a; } =20 - if (flags & float_muladd_negate_c) { - c->sign ^=3D 1; + if (unlikely(ab_mask =3D=3D float_cmask_infzero)) { + /* Inf * Zero =3D=3D NaN */ + float_raise(float_flag_invalid | float_flag_invalid_imz, s); + goto d_nan; } =20 - /* Compute the sign of the product into A. */ - a->sign ^=3D b->sign; - if (flags & float_muladd_negate_product) { - a->sign ^=3D 1; - } - - if (unlikely(!cmask_is_only_normals(ab_mask))) { - if (unlikely(ab_mask =3D=3D float_cmask_infzero)) { - float_raise(float_flag_invalid | float_flag_invalid_imz, s); + if (unlikely(ab_mask & float_cmask_inf)) { + if ((c_mask & float_cmask_inf) && p_sign !=3D c_sign) { + /* Inf - Inf =3D=3D NaN */ + float_raise(float_flag_invalid | float_flag_invalid_isi, s); goto d_nan; } - - if (ab_mask & float_cmask_inf) { - if (c->cls =3D=3D float_class_inf && a->sign !=3D c->sign) { - float_raise(float_flag_invalid | float_flag_invalid_isi, s= ); - goto d_nan; - } - goto return_inf; - } - - g_assert(ab_mask & float_cmask_zero); - if (is_anynorm(c->cls)) { - *a =3D *c; - goto finish_sign; - } - if (c->cls =3D=3D float_class_zero) { - if (flags & float_muladd_suppress_add_product_zero) { - a->sign =3D c->sign; - } else if (a->sign !=3D c->sign) { - goto return_sub_zero; - } - goto return_zero; - } - g_assert(c->cls =3D=3D float_class_inf); + /* Inf + C =3D=3D Inf */ + record_denormals_used(abc_mask, s); + a->sign =3D p_sign; + a->cls =3D float_class_inf; + return a; } - - if (unlikely(c->cls =3D=3D float_class_inf)) { - a->sign =3D c->sign; - goto return_inf; - } - - /* Perform the multiplication step. */ - p_widen.sign =3D a->sign; - p_widen.exp =3D a->exp + b->exp + 1; - fracN(mulw)(&p_widen, a, b); - if (!(p_widen.frac_hi & DECOMPOSED_IMPLICIT_BIT)) { - fracW(add)(&p_widen, &p_widen, &p_widen); - p_widen.exp -=3D 1; - } - - /* Perform the addition step. */ - if (c->cls !=3D float_class_zero) { - /* Zero-extend C to less significant bits. */ - fracN(widen)(&c_widen, c); - c_widen.exp =3D c->exp; - - if (a->sign =3D=3D c->sign) { - partsW(add_normal)(&p_widen, &c_widen); - } else if (!partsW(sub_normal)(&p_widen, &c_widen)) { - goto return_sub_zero; - } - } - - /* Narrow with sticky bit, for proper rounding later. */ - fracN(truncjam)(a, &p_widen); - a->sign =3D p_widen.sign; - a->exp =3D p_widen.exp; - - finish_sign: - /* - * All result types except for "return the default NaN - * because this is an Invalid Operation" go through here; - * this matches the set of cases where we consumed a - * denormal input. - */ record_denormals_used(abc_mask, s); - return a; + + /* Only remaining cases are zero product or inf addend. */ + assert((ab_mask & float_cmask_zero) | (c_mask & float_cmask_inf)); + + /* + * P + Inf =3D=3D Inf, or + * 0 + C =3D=3D C, + * except for 0 - 0, which needs special rounding, + * except for when we want to suppress this addition step. + */ + if (!(c_mask & float_cmask_zero) + || p_sign =3D=3D c_sign + || (flags & float_muladd_suppress_add_product_zero)) { + c->sign =3D c_sign; + return c; + } =20 return_sub_zero: + /* 0 - 0 =3D=3D -0 for round_down, +0 otherwise. */ a->sign =3D s->float_rounding_mode =3D=3D float_round_down; - return_zero: a->cls =3D float_class_zero; - goto finish_sign; - - return_inf: - a->cls =3D float_class_inf; - goto finish_sign; + return a; =20 d_nan: *a =3D partsN(default_nan)(s); --=20 2.43.0