From patchwork Wed Dec 19 22:56:24 2018 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Joseph Myers X-Patchwork-Id: 1016394 Return-Path: X-Original-To: incoming@patchwork.ozlabs.org Delivered-To: patchwork-incoming@bilbo.ozlabs.org Authentication-Results: ozlabs.org; spf=pass (mailfrom) smtp.mailfrom=sourceware.org (client-ip=209.132.180.131; helo=sourceware.org; envelope-from=libc-alpha-return-98620-incoming=patchwork.ozlabs.org@sourceware.org; receiver=) Authentication-Results: ozlabs.org; dmarc=none (p=none dis=none) header.from=codesourcery.com Authentication-Results: ozlabs.org; dkim=pass (1024-bit key; secure) header.d=sourceware.org header.i=@sourceware.org header.b="eevilgmh"; dkim-atps=neutral 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 43KqyL6nbDz9sD4 for ; Thu, 20 Dec 2018 09:56:42 +1100 (AEDT) DomainKey-Signature: a=rsa-sha1; c=nofws; d=sourceware.org; h=list-id :list-unsubscribe:list-subscribe:list-archive:list-post :list-help:sender:date:from:to:subject:message-id:mime-version :content-type; q=dns; s=default; b=fTF0m6u1ZExAcjB0TmHTUxlg8PUgf ongJK/P05uQRJYZB4UNUQUiKUHSY3va4YH8QB3bb2DOwcJNWM9URXfSLwpkx0lft lxwASSjv+EBNX8Q1fIgw/fUFzKUX9mRubiPrulwvWXDyiQASS756WoKxhDprmcJe s8NPyR9Xm7KItM= 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:date:from:to:subject:message-id:mime-version :content-type; s=default; bh=6wO50TRHgCaqHPg8RZk+pBZNKX4=; b=eev ilgmhNnHNGIlGHl+3iJ9fnEzi4/Evn4xXjaI++aqlocXj/RCgwnD2Or9AD0MRYqo L9um333BtlyheZL+ZUCq0RqwIH3qsvpNJ7gfesTE6zukf6o1k6b2ttt2Nabz9mT8 njjV9yxkrjTujkbZK7QWYK3gIMfoJ8zmKAbEdU34= Received: (qmail 13781 invoked by alias); 19 Dec 2018 22:56:36 -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 13757 invoked by uid 89); 19 Dec 2018 22:56:35 -0000 Authentication-Results: sourceware.org; auth=none X-Spam-SWARE-Status: No, score=-26.9 required=5.0 tests=BAYES_00, GIT_PATCH_0, GIT_PATCH_1, GIT_PATCH_2, GIT_PATCH_3, KAM_SHORT, RCVD_IN_DNSWL_NONE, SPF_PASS autolearn=ham version=3.3.2 spammy= X-HELO: relay1.mentorg.com Date: Wed, 19 Dec 2018 22:56:24 +0000 From: Joseph Myers To: Subject: Remove x86 mathinline.h [committed] Message-ID: User-Agent: Alpine 2.21 (DEB 202 2017-01-01) MIME-Version: 1.0 After previous cleanups, the only code in the x86 bits/mathinline.h that is relevant with current compilers is the inline of __ieee754_atan2l that is conditional on __LIBC_INTERNAL_MATH_INLINES (i.e. for when libm itself is being built). This inline is something that does belong in glibc not GCC, since __ieee754_atan2l is a purely internal function name. This patch moves that inline to a new sysdeps/x86/fpu/math_private.h, removing the bits/mathinline.h header. Note that previously the inline was only for non-SSE 32-bit x86. That condition does not make sense, however, for a long double function; if it's not inlined, exactly the same x87 instruction will end up getting used by the out-of-line function, for both 32-bit and 64-bit. So that condition is not retained in the new version. Tested for x86_64 and x86. As expected, installed stripped shared libraries are unchanged for 32-bit x86, but installed stripped libm.so is changed for x86_64 because calls to __ieee754_atan2l start being inlined where previously they were out of line calls. (The same change to start inlining the function would presumably also apply for 32-bit built with -mfpmath=sse, but that's not a configuration I've tested.) Committed. 2018-12-19 Joseph Myers * sysdeps/x86/fpu/math_private.h: New file. * sysdeps/x86/fpu/bits/mathinline.h: Remove. diff --git a/sysdeps/x86/fpu/bits/mathinline.h b/sysdeps/x86/fpu/bits/mathinline.h deleted file mode 100644 index 95ab9f5b82..0000000000 --- a/sysdeps/x86/fpu/bits/mathinline.h +++ /dev/null @@ -1,295 +0,0 @@ -/* Inline math functions for i387 and SSE. - Copyright (C) 1995-2018 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 _MATH_H -# error "Never use directly; include instead." -#endif - -#ifndef __extern_always_inline -# define __MATH_INLINE __inline -#else -# define __MATH_INLINE __extern_always_inline -#endif - -/* Disable x87 inlines when -fpmath=sse is passed and also when we're building - on x86_64. Older gcc (gcc-3.2 for example) does not define __SSE2_MATH__ - for x86_64. */ -#if !defined __SSE2_MATH__ && !defined __x86_64__ -# if ((!defined __NO_MATH_INLINES || defined __LIBC_INTERNAL_MATH_INLINES) \ - && defined __OPTIMIZE__) - -/* The inline functions do not set errno or raise necessarily the - correct exceptions. */ -# undef math_errhandling - -/* A macro to define float, double, and long double versions of various - math functions for the ix87 FPU. FUNC is the function name (which will - be suffixed with f and l for the float and long double version, - respectively). OP is the name of the FPU operation. - We define two sets of macros. The set with the additional NP - doesn't add a prototype declaration. */ - -# ifdef __USE_ISOC99 -# define __inline_mathop(func, op) \ - __inline_mathop_ (double, func, op) \ - __inline_mathop_ (float, __CONCAT(func,f), op) \ - __inline_mathop_ (long double, __CONCAT(func,l), op) -# define __inline_mathopNP(func, op) \ - __inline_mathopNP_ (double, func, op) \ - __inline_mathopNP_ (float, __CONCAT(func,f), op) \ - __inline_mathopNP_ (long double, __CONCAT(func,l), op) -# else -# define __inline_mathop(func, op) \ - __inline_mathop_ (double, func, op) -# define __inline_mathopNP(func, op) \ - __inline_mathopNP_ (double, func, op) -# endif - -# define __inline_mathop_(float_type, func, op) \ - __inline_mathop_decl_ (float_type, func, op, "0" (__x)) -# define __inline_mathopNP_(float_type, func, op) \ - __inline_mathop_declNP_ (float_type, func, op, "0" (__x)) - - -# ifdef __USE_ISOC99 -# define __inline_mathop_decl(func, op, params...) \ - __inline_mathop_decl_ (double, func, op, params) \ - __inline_mathop_decl_ (float, __CONCAT(func,f), op, params) \ - __inline_mathop_decl_ (long double, __CONCAT(func,l), op, params) -# define __inline_mathop_declNP(func, op, params...) \ - __inline_mathop_declNP_ (double, func, op, params) \ - __inline_mathop_declNP_ (float, __CONCAT(func,f), op, params) \ - __inline_mathop_declNP_ (long double, __CONCAT(func,l), op, params) -# else -# define __inline_mathop_decl(func, op, params...) \ - __inline_mathop_decl_ (double, func, op, params) -# define __inline_mathop_declNP(func, op, params...) \ - __inline_mathop_declNP_ (double, func, op, params) -# endif - -# define __inline_mathop_decl_(float_type, func, op, params...) \ - __MATH_INLINE float_type func (float_type) __THROW; \ - __inline_mathop_declNP_ (float_type, func, op, params) - -# define __inline_mathop_declNP_(float_type, func, op, params...) \ - __MATH_INLINE float_type __NTH (func (float_type __x)) \ - { \ - register float_type __result; \ - __asm __volatile__ (op : "=t" (__result) : params); \ - return __result; \ - } - - -# ifdef __USE_ISOC99 -# define __inline_mathcode(func, arg, code) \ - __inline_mathcode_ (double, func, arg, code) \ - __inline_mathcode_ (float, __CONCAT(func,f), arg, code) \ - __inline_mathcode_ (long double, __CONCAT(func,l), arg, code) -# define __inline_mathcodeNP(func, arg, code) \ - __inline_mathcodeNP_ (double, func, arg, code) \ - __inline_mathcodeNP_ (float, __CONCAT(func,f), arg, code) \ - __inline_mathcodeNP_ (long double, __CONCAT(func,l), arg, code) -# define __inline_mathcode2(func, arg1, arg2, code) \ - __inline_mathcode2_ (double, func, arg1, arg2, code) \ - __inline_mathcode2_ (float, __CONCAT(func,f), arg1, arg2, code) \ - __inline_mathcode2_ (long double, __CONCAT(func,l), arg1, arg2, code) -# define __inline_mathcodeNP2(func, arg1, arg2, code) \ - __inline_mathcodeNP2_ (double, func, arg1, arg2, code) \ - __inline_mathcodeNP2_ (float, __CONCAT(func,f), arg1, arg2, code) \ - __inline_mathcodeNP2_ (long double, __CONCAT(func,l), arg1, arg2, code) -# define __inline_mathcode3(func, arg1, arg2, arg3, code) \ - __inline_mathcode3_ (double, func, arg1, arg2, arg3, code) \ - __inline_mathcode3_ (float, __CONCAT(func,f), arg1, arg2, arg3, code) \ - __inline_mathcode3_ (long double, __CONCAT(func,l), arg1, arg2, arg3, code) -# define __inline_mathcodeNP3(func, arg1, arg2, arg3, code) \ - __inline_mathcodeNP3_ (double, func, arg1, arg2, arg3, code) \ - __inline_mathcodeNP3_ (float, __CONCAT(func,f), arg1, arg2, arg3, code) \ - __inline_mathcodeNP3_ (long double, __CONCAT(func,l), arg1, arg2, arg3, code) -# else -# define __inline_mathcode(func, arg, code) \ - __inline_mathcode_ (double, func, (arg), code) -# define __inline_mathcodeNP(func, arg, code) \ - __inline_mathcodeNP_ (double, func, (arg), code) -# define __inline_mathcode2(func, arg1, arg2, code) \ - __inline_mathcode2_ (double, func, arg1, arg2, code) -# define __inline_mathcodeNP2(func, arg1, arg2, code) \ - __inline_mathcodeNP2_ (double, func, arg1, arg2, code) -# define __inline_mathcode3(func, arg1, arg2, arg3, code) \ - __inline_mathcode3_ (double, func, arg1, arg2, arg3, code) -# define __inline_mathcodeNP3(func, arg1, arg2, arg3, code) \ - __inline_mathcodeNP3_ (double, func, arg1, arg2, arg3, code) -# endif - -# define __inline_mathcode_(float_type, func, arg, code) \ - __MATH_INLINE float_type func (float_type) __THROW; \ - __inline_mathcodeNP_(float_type, func, arg, code) - -# define __inline_mathcodeNP_(float_type, func, arg, code) \ - __MATH_INLINE float_type __NTH (func (float_type arg)) \ - { \ - code; \ - } - - -# define __inline_mathcode2_(float_type, func, arg1, arg2, code) \ - __MATH_INLINE float_type func (float_type, float_type) __THROW; \ - __inline_mathcodeNP2_ (float_type, func, arg1, arg2, code) - -# define __inline_mathcodeNP2_(float_type, func, arg1, arg2, code) \ - __MATH_INLINE float_type __NTH (func (float_type arg1, float_type arg2)) \ - { \ - code; \ - } - -# define __inline_mathcode3_(float_type, func, arg1, arg2, arg3, code) \ - __MATH_INLINE float_type func (float_type, float_type, float_type) __THROW; \ - __inline_mathcodeNP3_(float_type, func, arg1, arg2, arg3, code) - -# define __inline_mathcodeNP3_(float_type, func, arg1, arg2, arg3, code) \ - __MATH_INLINE float_type __NTH (func (float_type arg1, float_type arg2, \ - float_type arg3)) \ - { \ - code; \ - } -# endif - - -# if !defined __NO_MATH_INLINES && defined __OPTIMIZE__ -/* Miscellaneous functions */ - -/* __FAST_MATH__ is defined by gcc -ffast-math. */ -# ifdef __FAST_MATH__ -/* Optimized inline implementation, sometimes with reduced precision - and/or argument range. */ - -# if __GNUC_PREREQ (3, 5) -# define __expm1_code \ - register long double __temp; \ - __temp = __builtin_expm1l (__x); \ - return __temp ? __temp : __x -# else -# define __expm1_code \ - register long double __value; \ - register long double __exponent; \ - register long double __temp; \ - __asm __volatile__ \ - ("fldl2e # e^x - 1 = 2^(x * log2(e)) - 1\n\t" \ - "fmul %%st(1) # x * log2(e)\n\t" \ - "fst %%st(1)\n\t" \ - "frndint # int(x * log2(e))\n\t" \ - "fxch\n\t" \ - "fsub %%st(1) # fract(x * log2(e))\n\t" \ - "f2xm1 # 2^(fract(x * log2(e))) - 1\n\t" \ - "fscale # 2^(x * log2(e)) - 2^(int(x * log2(e)))\n\t" \ - : "=t" (__value), "=u" (__exponent) : "0" (__x)); \ - __asm __volatile__ \ - ("fscale # 2^int(x * log2(e))\n\t" \ - : "=t" (__temp) : "0" (1.0), "u" (__exponent)); \ - __temp -= 1.0; \ - __temp += __value; \ - return __temp ? __temp : __x -# endif -__inline_mathcodeNP_ (long double, __expm1l, __x, __expm1_code) - -# if __GNUC_PREREQ (3, 4) -__inline_mathcodeNP_ (long double, __expl, __x, return __builtin_expl (__x)) -# else -# define __exp_code \ - register long double __value; \ - register long double __exponent; \ - __asm __volatile__ \ - ("fldl2e # e^x = 2^(x * log2(e))\n\t" \ - "fmul %%st(1) # x * log2(e)\n\t" \ - "fst %%st(1)\n\t" \ - "frndint # int(x * log2(e))\n\t" \ - "fxch\n\t" \ - "fsub %%st(1) # fract(x * log2(e))\n\t" \ - "f2xm1 # 2^(fract(x * log2(e))) - 1\n\t" \ - : "=t" (__value), "=u" (__exponent) : "0" (__x)); \ - __value += 1.0; \ - __asm __volatile__ \ - ("fscale" \ - : "=t" (__value) : "0" (__value), "u" (__exponent)); \ - return __value -__inline_mathcodeNP (exp, __x, __exp_code) -__inline_mathcodeNP_ (long double, __expl, __x, __exp_code) -# endif -# endif /* __FAST_MATH__ */ - - -# ifdef __FAST_MATH__ -# if !__GNUC_PREREQ (3,3) -__inline_mathopNP (sqrt, "fsqrt") -__inline_mathopNP_ (long double, __sqrtl, "fsqrt") -# define __libc_sqrtl(n) __sqrtl (n) -# else -# define __libc_sqrtl(n) __builtin_sqrtl (n) -# endif -# endif - -# if __GNUC_PREREQ (2, 8) -__inline_mathcodeNP_ (double, fabs, __x, return __builtin_fabs (__x)) -# ifdef __USE_ISOC99 -__inline_mathcodeNP_ (float, fabsf, __x, return __builtin_fabsf (__x)) -__inline_mathcodeNP_ (long double, fabsl, __x, return __builtin_fabsl (__x)) -# endif -__inline_mathcodeNP_ (long double, __fabsl, __x, return __builtin_fabsl (__x)) -# else -__inline_mathop (fabs, "fabs") -__inline_mathop_ (long double, __fabsl, "fabs") -# endif - -__inline_mathcode_ (long double, __sgn1l, __x, \ - __extension__ union { long double __xld; unsigned int __xi[3]; } __n = \ - { __xld: __x }; \ - __n.__xi[2] = (__n.__xi[2] & 0x8000) | 0x3fff; \ - __n.__xi[1] = 0x80000000; \ - __n.__xi[0] = 0; \ - return __n.__xld) - - -/* Optimized versions for some non-standardized functions. */ -# ifdef __USE_ISOC99 - -# ifdef __FAST_MATH__ -__inline_mathcodeNP (expm1, __x, __expm1_code) - -# endif -# endif - - -/* Undefine some of the large macros which are not used anymore. */ -# ifdef __FAST_MATH__ -# undef __expm1_code -# undef __exp_code -# endif /* __FAST_MATH__ */ - -# endif /* __NO_MATH_INLINES */ - - -/* This code is used internally in the GNU libc. */ -# ifdef __LIBC_INTERNAL_MATH_INLINES -__inline_mathcode2_ (long double, __ieee754_atan2l, __y, __x, - register long double __value; - __asm __volatile__ ("fpatan\n\t" - : "=t" (__value) - : "0" (__x), "u" (__y) : "st(1)"); - return __value;) -# endif - -#endif /* !__SSE2_MATH__ && !__x86_64__ */ diff --git a/sysdeps/x86/fpu/math_private.h b/sysdeps/x86/fpu/math_private.h new file mode 100644 index 0000000000..dba29d6428 --- /dev/null +++ b/sysdeps/x86/fpu/math_private.h @@ -0,0 +1,32 @@ +/* Private inline math functions for x86. + Copyright (C) 1995-2018 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 X86_MATH_PRIVATE_H +#define X86_MATH_PRIVATE_H 1 + +#include_next + +__extern_always_inline long double +__NTH (__ieee754_atan2l (long double y, long double x)) +{ + long double ret; + __asm__ __volatile__ ("fpatan" : "=t" (ret) : "0" (x), "u" (y) : "st(1)"); + return ret; +} + +#endif