From patchwork Tue Aug 30 23:39:34 2016 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: "Paul E. Murphy" X-Patchwork-Id: 664327 Return-Path: X-Original-To: incoming@patchwork.ozlabs.org Delivered-To: patchwork-incoming@bilbo.ozlabs.org Received: from sourceware.org (server1.sourceware.org [209.132.180.131]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by ozlabs.org (Postfix) with ESMTPS id 3sP4l86XmYz9s8x for ; Wed, 31 Aug 2016 09:40:36 +1000 (AEST) Authentication-Results: ozlabs.org; dkim=pass (1024-bit key; secure) header.d=sourceware.org header.i=@sourceware.org header.b=Prf34IDU; dkim-atps=neutral DomainKey-Signature: a=rsa-sha1; c=nofws; d=sourceware.org; h=list-id :list-unsubscribe:list-subscribe:list-archive:list-post :list-help:sender:from:to:subject:date:message-id; q=dns; s= default; b=fpsh6wdHLjI/4utPZjbcrn3giOLGfn79KBrryrww8TkaKe9MQ4BaX Flh8NRHXKsjgzhR6xnxECe+mk0FXrP0znPcb/0v6YsMjvZNgHCObZnBRlPJ6rqnC 7Mw+uDsfviZLm1qSMW0byZJ9Z717z+a0zYXWtx0GgnCOFAvZKJdsFM= DKIM-Signature: v=1; a=rsa-sha1; c=relaxed; d=sourceware.org; h=list-id :list-unsubscribe:list-subscribe:list-archive:list-post :list-help:sender:from:to:subject:date:message-id; s=default; bh=65y+qh/cyKkhjChaMMtvUq7ecdc=; b=Prf34IDUMsUcZ5Z0qRM4+G+adT6M iGTlugxi1tuPDSX9nktNTjF8zLW887jBHHVB+Z9wF5TD9Bto+MZr7ezAg49TVzpO FxKByCS8Moa0f43lnmjK7ucven7YVdt7lgWWEMIK68IPz5YdLbxS48R6NetNCWwE Ivfv5ZGrnVZS+U4= Received: (qmail 35271 invoked by alias); 30 Aug 2016 23:39:56 -0000 Mailing-List: contact libc-alpha-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Unsubscribe: List-Subscribe: List-Archive: List-Post: List-Help: , Sender: libc-alpha-owner@sourceware.org Delivered-To: mailing list libc-alpha@sourceware.org Received: (qmail 35021 invoked by uid 89); 30 Aug 2016 23:39:54 -0000 Authentication-Results: sourceware.org; auth=none X-Virus-Found: No X-Spam-SWARE-Status: No, score=-1.1 required=5.0 tests=AWL, BAYES_00, KAM_LAZY_DOMAIN_SECURITY, RCVD_IN_DNSWL_LOW, RCVD_IN_SEMBACKSCATTER autolearn=no version=3.3.2 spammy=drepper, __set_errno, UD:math_ldbl_opt.h, math_ldbl_opth X-HELO: mx0a-001b2d01.pphosted.com X-IBM-Helo: d03dlp03.boulder.ibm.com X-IBM-MailFrom: murphyp@linux.vnet.ibm.com From: "Paul E. Murphy" To: libc-alpha@sourceware.org Subject: [PATCH 1/5] Make common fdim implementation generic. Date: Tue, 30 Aug 2016 18:39:34 -0500 X-TM-AS-GCONF: 00 X-Content-Scanned: Fidelis XPS MAILER x-cbid: 16083023-0024-0000-0000-0000147168DA X-IBM-SpamModules-Scores: X-IBM-SpamModules-Versions: BY=3.00005680; HX=3.00000240; KW=3.00000007; PH=3.00000004; SC=3.00000184; SDB=6.00751825; UDB=6.00355389; IPR=6.00524631; BA=6.00004681; NDR=6.00000001; ZLA=6.00000005; ZF=6.00000009; ZB=6.00000000; ZP=6.00000000; ZH=6.00000000; ZU=6.00000002; MB=3.00012523; XFM=3.00000011; UTC=2016-08-30 23:39:40 X-IBM-AV-DETECTION: SAVI=unused REMOTE=unused XFE=unused x-cbparentid: 16083023-0025-0000-0000-00004408670B Message-Id: X-Proofpoint-Virus-Version: vendor=fsecure engine=2.50.10432:, , definitions=2016-08-30_10:, , signatures=0 X-Proofpoint-Spam-Details: rule=outbound_notspam policy=outbound score=0 spamscore=0 suspectscore=4 malwarescore=0 phishscore=0 adultscore=0 bulkscore=0 classifier=spam adjust=0 reason=mlx scancount=1 engine=8.0.1-1604210000 definitions=main-1608300225 The only difference is the usage of math_narrow_eval when building s_fdiml.c. This should be harmless for long double, but I did observe some code generation changes on m68k, but lack the resources to test it. Likewise, to more easily support overriding symbol generation, the aliasing macros are always conditionally defined on their absence to reduce boilerplate. I also ran builds for i486, ppc64, sparcv9, aarch64, s390x and observed no changes to s_fdim* objects. * math/Makefile (gen-libm-calls): Add s_fdim. (libm-calls): Move to above. * math/s_fdim.c: Renamed and refactored into to ... * math/s_fdim_template.c: New file. * math/s_fdiml.c: Removed. * sysdeps/ieee754/ldbl-opt/s_fdim.c: Removed. * sysdeps/ieee754/ldbl-opt/s_fdiml.c: Removed. * sysdeps/generic/math-type-macros-double.h: (declare_mgen_alias): Don't declare if already declared. * sysdeps/ieee754/ldbl-opt/math-type-macros-ldouble.h: Likewise. * sysdeps/ieee754/ldbl-opt/math-type-macros-double.h: (declare_mgen_alias): Don't declare if already declared. (M_LIBM_NEED_COMPAT): Likewise. (declare_mgen_libm_compat): Likewise. (LDOUBLE_fdiml_libm_version): New macro. * sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_fdim.c: Include math/s_fdim.c now that ldbl-opt version is not needed. (declare_mgen_alias): New macro to disable aliasing. * sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_fdimf-vis3.c: (declare_mgen_alias): Likewise. * sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_fdimf.c: (declare_mgen_alias): Likewise. * sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_fdim-vis3.c: (declare_mgen_alias): Likewise. (M_LIBM_NEED_COMPAT): Likewise. --- math/Makefile | 4 +-- math/s_fdim.c | 42 ---------------------- math/s_fdim_template.c | 40 +++++++++++++++++++++ math/s_fdimf.c | 38 -------------------- math/s_fdiml.c | 35 ------------------ sysdeps/generic/math-type-macros-double.h | 4 +-- sysdeps/ieee754/ldbl-opt/math-type-macros-double.h | 13 ++++--- .../ieee754/ldbl-opt/math-type-macros-ldouble.h | 6 ++-- sysdeps/ieee754/ldbl-opt/s_fdim.c | 5 --- sysdeps/ieee754/ldbl-opt/s_fdiml.c | 5 --- .../sparc32/sparcv9/fpu/multiarch/s_fdim-vis3.c | 2 ++ .../sparc/sparc32/sparcv9/fpu/multiarch/s_fdim.c | 3 +- .../sparc32/sparcv9/fpu/multiarch/s_fdimf-vis3.c | 1 + .../sparc/sparc32/sparcv9/fpu/multiarch/s_fdimf.c | 2 ++ 14 files changed, 64 insertions(+), 136 deletions(-) delete mode 100644 math/s_fdim.c create mode 100644 math/s_fdim_template.c delete mode 100644 math/s_fdimf.c delete mode 100644 math/s_fdiml.c delete mode 100644 sysdeps/ieee754/ldbl-opt/s_fdim.c delete mode 100644 sysdeps/ieee754/ldbl-opt/s_fdiml.c diff --git a/math/Makefile b/math/Makefile index f1b7937..d4b6d98 100644 --- a/math/Makefile +++ b/math/Makefile @@ -49,7 +49,7 @@ gen-libm-calls = cargF conjF cimagF crealF cabsF s_cacosF \ s_cacoshF s_ccosF s_ccoshF s_casinF s_csinF s_casinhF \ k_casinhF s_csinhF k_casinhF s_csinhF s_catanhF s_catanF \ s_ctanF s_ctanhF s_cexpF s_clogF s_cprojF s_csqrtF \ - s_cpowF s_clog10F + s_cpowF s_clog10F s_fdimF libm-calls = \ e_acosF e_acoshF e_asinF e_atan2F e_atanhF e_coshF e_expF e_fmodF \ @@ -65,7 +65,7 @@ libm-calls = \ w_tgammaF w_hypotF w_j0F w_j1F w_jnF w_lgammaF w_lgammaF_r \ w_logF w_log10F w_powF w_remainderF w_scalbF w_sinhF w_sqrtF \ w_ilogbF \ - s_fpclassifyF s_fmaxF s_fminF s_fdimF s_nanF s_truncF \ + s_fpclassifyF s_fmaxF s_fminF s_nanF s_truncF \ s_remquoF e_log2F e_exp2F s_roundF s_nearbyintF s_sincosF \ s_fmaF s_lrintF s_llrintF s_lroundF s_llroundF e_exp10F w_log2F \ s_issignalingF $(calls:s_%=m_%) x2y2m1F \ diff --git a/math/s_fdim.c b/math/s_fdim.c deleted file mode 100644 index 1786521..0000000 --- a/math/s_fdim.c +++ /dev/null @@ -1,42 +0,0 @@ -/* Return positive difference between arguments. - Copyright (C) 1997-2016 Free Software Foundation, Inc. - This file is part of the GNU C Library. - Contributed by Ulrich Drepper , 1997. - - 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 -#include -#include - -double -__fdim (double x, double y) -{ - if (islessequal (x, y)) - return 0.0; - - double r = math_narrow_eval (x - y); - if (isinf (r) && !isinf (x) && !isinf (y)) - __set_errno (ERANGE); - - return r; -} -#ifndef __fdim -weak_alias (__fdim, fdim) -# ifdef NO_LONG_DOUBLE -strong_alias (__fdim, __fdiml) -weak_alias (__fdim, fdiml) -# endif -#endif diff --git a/math/s_fdim_template.c b/math/s_fdim_template.c new file mode 100644 index 0000000..d3e7379 --- /dev/null +++ b/math/s_fdim_template.c @@ -0,0 +1,40 @@ +/* Return positive difference between arguments. + Copyright (C) 1997-2016 Free Software Foundation, Inc. + This file is part of the GNU C Library. + Contributed by Ulrich Drepper , 1997. + + 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 +#include +#include + +FLOAT +M_DECL_FUNC (__fdim) (FLOAT x, FLOAT y) +{ + if (islessequal (x, y)) + return 0; + + FLOAT r = math_narrow_eval (x - y); + if (isinf (r) && !isinf (x) && !isinf (y)) + __set_errno (ERANGE); + + return r; +} +declare_mgen_alias (__fdim, fdim); + +#if M_LIBM_NEED_COMPAT (fdim) +declare_mgen_libm_compat (__fdim, fdim) +#endif diff --git a/math/s_fdimf.c b/math/s_fdimf.c deleted file mode 100644 index e457f9d..0000000 --- a/math/s_fdimf.c +++ /dev/null @@ -1,38 +0,0 @@ -/* Return positive difference between arguments. - Copyright (C) 1997-2016 Free Software Foundation, Inc. - This file is part of the GNU C Library. - Contributed by Ulrich Drepper , 1997. - - 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 -#include -#include - -float -__fdimf (float x, float y) -{ - if (islessequal (x, y)) - return 0.0f; - - float r = math_narrow_eval (x - y); - if (isinf (r) && !isinf (x) && !isinf (y)) - __set_errno (ERANGE); - - return r; -} -#ifndef __fdimf -weak_alias (__fdimf, fdimf) -#endif diff --git a/math/s_fdiml.c b/math/s_fdiml.c deleted file mode 100644 index 4a1f672..0000000 --- a/math/s_fdiml.c +++ /dev/null @@ -1,35 +0,0 @@ -/* Return positive difference between arguments. - Copyright (C) 1997-2016 Free Software Foundation, Inc. - This file is part of the GNU C Library. - Contributed by Ulrich Drepper , 1997. - - 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 -#include - -long double -__fdiml (long double x, long double y) -{ - if (islessequal (x, y)) - return 0.0f; - - long double r = x - y; - if (isinf (r) && !isinf (x) && !isinf (y)) - __set_errno (ERANGE); - - return r; -} -weak_alias (__fdiml, fdiml) diff --git a/sysdeps/generic/math-type-macros-double.h b/sysdeps/generic/math-type-macros-double.h index 4149b39..154f416 100644 --- a/sysdeps/generic/math-type-macros-double.h +++ b/sysdeps/generic/math-type-macros-double.h @@ -29,12 +29,12 @@ /* Machines without a distinct long double type alias long double functions to their double equivalent. */ -#if defined NO_LONG_DOUBLE +#if defined NO_LONG_DOUBLE && !defined declare_mgen_alias # define declare_mgen_alias(from, to) \ weak_alias (from, to) \ strong_alias (from, from ## l) \ weak_alias (from, to ## l) -#else +#elif !defined declare_mgen_alias # define declare_mgen_alias(from, to) \ weak_alias (M_SUF (from), M_SUF (to)) #endif diff --git a/sysdeps/ieee754/ldbl-opt/math-type-macros-double.h b/sysdeps/ieee754/ldbl-opt/math-type-macros-double.h index 8cb5694..e9b1508 100644 --- a/sysdeps/ieee754/ldbl-opt/math-type-macros-double.h +++ b/sysdeps/ieee754/ldbl-opt/math-type-macros-double.h @@ -44,15 +44,20 @@ #define LDOUBLE_cpowl_libm_version GLIBC_2_1 #define LDOUBLE_clog10l_libm_version GLIBC_2_1 #define LDOUBLE___clog10l_libm_version GLIBC_2_1 +#define LDOUBLE_fdiml_libm_version GLIBC_2_1 /* Define compat symbols for long double on platforms where it was not always a distinct type. */ -#define M_LIBM_NEED_COMPAT(f) \ - LONG_DOUBLE_COMPAT (libm, LDOUBLE_ ## f ## l_libm_version) +#if !defined M_LIBM_NEED_COMPAT +# define M_LIBM_NEED_COMPAT(f) \ + LONG_DOUBLE_COMPAT (libm, LDOUBLE_ ## f ## l_libm_version) +#endif -#define declare_mgen_libm_compat(from, to) \ - compat_symbol (libm, from, to ## l, \ +#if !defined declare_mgen_libm_compat +# define declare_mgen_libm_compat(from, to) \ + compat_symbol (libm, from, to ## l, \ LDOUBLE_ ## to ## l_libm_version); +#endif #include_next #endif diff --git a/sysdeps/ieee754/ldbl-opt/math-type-macros-ldouble.h b/sysdeps/ieee754/ldbl-opt/math-type-macros-ldouble.h index d2af4bb..30cb2fb 100644 --- a/sysdeps/ieee754/ldbl-opt/math-type-macros-ldouble.h +++ b/sysdeps/ieee754/ldbl-opt/math-type-macros-ldouble.h @@ -22,8 +22,10 @@ /* Use properly versioned symbols for long double on platforms where it was not always a distinct type. */ -#define declare_mgen_alias(from, to) \ - long_double_symbol (libm, from ## l, to ## l); +#if !defined declare_mgen_alias +# define declare_mgen_alias(from, to) \ + long_double_symbol (libm, from ## l, to ## l); +#endif #include_next #endif diff --git a/sysdeps/ieee754/ldbl-opt/s_fdim.c b/sysdeps/ieee754/ldbl-opt/s_fdim.c deleted file mode 100644 index 02c95bf..0000000 --- a/sysdeps/ieee754/ldbl-opt/s_fdim.c +++ /dev/null @@ -1,5 +0,0 @@ -#include -#include -#if LONG_DOUBLE_COMPAT(libm, GLIBC_2_1) -compat_symbol (libm, __fdim, fdiml, GLIBC_2_1); -#endif diff --git a/sysdeps/ieee754/ldbl-opt/s_fdiml.c b/sysdeps/ieee754/ldbl-opt/s_fdiml.c deleted file mode 100644 index 06b760b..0000000 --- a/sysdeps/ieee754/ldbl-opt/s_fdiml.c +++ /dev/null @@ -1,5 +0,0 @@ -#include -#undef weak_alias -#define weak_alias(n,a) -#include -long_double_symbol (libm, __fdiml, fdiml); diff --git a/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_fdim-vis3.c b/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_fdim-vis3.c index 2973b49..6176517 100644 --- a/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_fdim-vis3.c +++ b/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_fdim-vis3.c @@ -19,5 +19,7 @@ #include #define __fdim __fdim_vis3 +#define declare_mgen_alias(t, f) +#define M_LIBM_NEED_COMPAT(f) 0 #include diff --git a/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_fdim.c b/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_fdim.c index 9666741..2d07f31 100644 --- a/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_fdim.c +++ b/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_fdim.c @@ -27,6 +27,7 @@ sparc_libm_ifunc(__fdim, hwcap & HWCAP_SPARC_VIS3 ? __fdim_vis3 : __fdim_generic weak_alias (__fdim, fdim) # define __fdim __fdim_generic +# define declare_mgen_alias(t, f) #endif -#include +#include diff --git a/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_fdimf-vis3.c b/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_fdimf-vis3.c index 75997c6..3f35b0d 100644 --- a/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_fdimf-vis3.c +++ b/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_fdimf-vis3.c @@ -19,5 +19,6 @@ #include #define __fdimf __fdimf_vis3 +#define declare_mgen_alias(t, f) #include diff --git a/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_fdimf.c b/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_fdimf.c index 767520f..b9add1c 100644 --- a/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_fdimf.c +++ b/sysdeps/sparc/sparc32/sparcv9/fpu/multiarch/s_fdimf.c @@ -27,6 +27,8 @@ sparc_libm_ifunc(__fdimf, hwcap & HWCAP_SPARC_VIS3 ? __fdimf_vis3 : __fdimf_gene weak_alias (__fdimf, fdimf) # define __fdimf __fdimf_generic +# define declare_mgen_alias(t, f) + #endif #include