From owner-freebsd-standards@FreeBSD.ORG Mon Jan 18 11:07:07 2010 Return-Path: Delivered-To: freebsd-standards@FreeBSD.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id E8A16106566B for ; Mon, 18 Jan 2010 11:07:06 +0000 (UTC) (envelope-from owner-bugmaster@FreeBSD.org) Received: from freefall.freebsd.org (freefall.freebsd.org [IPv6:2001:4f8:fff6::28]) by mx1.freebsd.org (Postfix) with ESMTP id D6C518FC20 for ; Mon, 18 Jan 2010 11:07:06 +0000 (UTC) Received: from freefall.freebsd.org (localhost [127.0.0.1]) by freefall.freebsd.org (8.14.3/8.14.3) with ESMTP id o0IB76Sx047679 for ; Mon, 18 Jan 2010 11:07:06 GMT (envelope-from owner-bugmaster@FreeBSD.org) Received: (from gnats@localhost) by freefall.freebsd.org (8.14.3/8.14.3/Submit) id o0IB76we047677 for freebsd-standards@FreeBSD.org; Mon, 18 Jan 2010 11:07:06 GMT (envelope-from owner-bugmaster@FreeBSD.org) Date: Mon, 18 Jan 2010 11:07:06 GMT Message-Id: <201001181107.o0IB76we047677@freefall.freebsd.org> X-Authentication-Warning: freefall.freebsd.org: gnats set sender to owner-bugmaster@FreeBSD.org using -f From: FreeBSD bugmaster To: freebsd-standards@FreeBSD.org Cc: Subject: Current problem reports assigned to freebsd-standards@FreeBSD.org X-BeenThere: freebsd-standards@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Standards compliance List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 18 Jan 2010 11:07:07 -0000 Note: to view an individual PR, use: http://www.freebsd.org/cgi/query-pr.cgi?pr=(number). The following is a listing of current problems submitted by FreeBSD users. These represent problem reports covering all versions including experimental development code and obsolete releases. S Tracker Resp. Description -------------------------------------------------------------------------------- o stand/142803 standards j0 Bessel function inaccurate near zeros of the functi o stand/142255 standards scandir prototype in dirent.h isn't compliant with POS s stand/141705 standards [libc] [request] libc lacks cexp (and friends) o stand/135307 standards Boot Loader problem on Acer Aspire 5735 o stand/130067 standards Wrong numeric limits in system headers? o stand/129524 standards FreeBSD 7.0 isnt detecting my hardrives with raid5 o stand/129196 standards Inconsistent errno in strtol() o bin/125855 standards sh(1) allows for multiline, non-escaped control struct o stand/124860 standards flockfile(3) doesn't work when the memory has been exh o stand/123688 standards POSIX standard changes in unistd.h and grp.h o stand/121921 standards [patch] Add leap second support to at(1), atrun(8) o stand/121568 standards [patch] ln(1): wrong "ln -s" behaviour o stand/120947 standards xsm ignores system.xsm and .xsmstartup o stand/116826 standards [patch] sh support for POSIX character classes o stand/116477 standards rm(1): rm behaves unexpectedly when using -r and relat o bin/116413 standards incorrect getconf(1) handling of unsigned constants gi o stand/116081 standards make does not work with the directive sinclude p stand/107561 standards [libc] [patch] [request] Missing SUS function tcgetsid o stand/104743 standards [headers] [patch] Wrong values for _POSIX_ minimal lim o stand/100017 standards [Patch] Add fuser(1) functionality to fstat(1) o stand/96236 standards [patch] [posix] sed(1) incorrectly describes a functio o stand/96016 standards [headers] clock_getres et al should be in o stand/94729 standards [libc] fcntl() throws undocumented ENOTTY o kern/93705 standards [headers] [patch] ENODATA and EGREGIOUS (for glibc com o stand/92362 standards [headers] [patch] Missing SIGPOLL in kernel headers a stand/86484 standards [patch] mkfifo(1) uses wrong permissions o stand/83845 standards [libm] [patch] add log2() and log2f() support for libm o stand/82654 standards C99 long double math functions are missing o stand/81287 standards [patch] fingerd(8) might send a line not ending in CRL a stand/80293 standards sysconf() does not support well-defined unistd values o stand/79056 standards [feature request] [atch] regex(3) regression tests o stand/70813 standards [patch] ls(1) not Posix compliant o stand/66357 standards make POSIX conformance problem ('sh -e' & '+' command- s kern/64875 standards [libc] [patch] [request] add a system call: fdatasync( s stand/62858 standards malloc(0) not C99 compliant o stand/56476 standards cd9660 unicode support simple hack o stand/54839 standards [pcvt] pcvt deficits o stand/54833 standards [pcvt] more pcvt deficits o stand/54410 standards one-true-awk not POSIX compliant (no extended REs) o stand/46119 standards Priority problems for SCHED_OTHER using pthreads o stand/44425 standards getcwd() succeeds even if current dir has perm 000. p stand/41576 standards POSIX compliance of ln(1) o stand/39256 standards snprintf/vsnprintf aren't POSIX-conformant for strings s stand/36076 standards Implementation of POSIX fuser command o kern/27835 standards [libc] execve() doesn't conform to execve(2) spec in s a docs/26003 standards getgroups(2) lists NGROUPS_MAX but not syslimits.h s stand/24590 standards timezone function not compatible witn Single Unix Spec o bin/24390 standards ln(1) Replacing old dir-symlinks when using /bin/ln o stand/21519 standards sys/dir.h should be deprecated some more s bin/14925 standards getsubopt isn't poisonous enough 50 problems total. From owner-freebsd-standards@FreeBSD.ORG Tue Jan 19 01:50:47 2010 Return-Path: Delivered-To: freebsd-standards@FreeBSD.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id A3FFF1065670 for ; Tue, 19 Jan 2010 01:50:47 +0000 (UTC) (envelope-from sgk@troutmask.apl.washington.edu) Received: from troutmask.apl.washington.edu (troutmask.apl.washington.edu [128.208.78.105]) by mx1.freebsd.org (Postfix) with ESMTP id 84BA08FC08 for ; Tue, 19 Jan 2010 01:50:47 +0000 (UTC) Received: from troutmask.apl.washington.edu (localhost.apl.washington.edu [127.0.0.1]) by troutmask.apl.washington.edu (8.14.3/8.14.3) with ESMTP id o0J1orZM045327 for ; Mon, 18 Jan 2010 17:50:53 -0800 (PST) (envelope-from sgk@troutmask.apl.washington.edu) Received: (from sgk@localhost) by troutmask.apl.washington.edu (8.14.3/8.14.3/Submit) id o0J1or1C045326 for freebsd-standards@FreeBSD.org; Mon, 18 Jan 2010 17:50:53 -0800 (PST) (envelope-from sgk) Date: Mon, 18 Jan 2010 17:50:53 -0800 From: Steve Kargl To: freebsd-standards@FreeBSD.org Message-ID: <20100119015053.GA45200@troutmask.apl.washington.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.4.2.3i Cc: Subject: [PATCH] implementation of expl(3) X-BeenThere: freebsd-standards@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Standards compliance List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 19 Jan 2010 01:50:47 -0000 I've finally found the last bug in my implementation of Tang's table-driven algorithm for computing expl(x). The following code has been tested on i686 and amd64 class hardware, giving the following: 0n i686-class hardware with fpsetprec(FP_PE). Values tested: 2383852701 Max ULP: 0.506987 On amd64 hardware. Values tested: 2383852701 Max ULP: 0.507022 Tested values are in the range LDBL_MIN_EXP * LN2 <= x <= LDBL_MAX_EXP * LN2 where x is a float, and to populate the lower order bits I did x = (x / (long double) M_PI) * M_PI prior to calls to expl(x). This code belongs in msun/ld80. I do not have ld128 hardware, so I haven't tried to produce a e_expl.c file this precision. PS: if someone wants the float and double versions of the table-driven expf() and exp(), I have those implemented as well with the caveat that exceptional values are tested via isinf() and isnan(). -- Steve /* * Implemenation of expl(x) in Intel 80-bit long double format. * The significand is 64 bits. This is based on * * PTP Tang, "Table-driven implementation of the exponential function * in IEEE floating-point arithmetic," ACM Trans. Math. Soft., 15, * 144-157 (1989). * * where the 32 table entries has been expanded to NUM (see below). * */ #include #include #include "fpmath.h" #define BIAS (LDBL_MAX_EXP - 1) /* * The range of overflow to +Inf and of underflow to +0 is given by * 41021 * ln(2). 41021 is obtained from -16382-64+1+24576 where * -16382 is the minimum exponent, 64 is the precision, and 24576 is * a bias adjustment given in IEEE 754 as 3*2^(e-2) with e being the * width of the exponent in bits. */ static const long double THRESHOLD_1 = 2.84335904937495165381e+04L, /* 41021 * ln(2) */ THRESHOLD_2 = 2.71050543121376108502e-20L, /* 2^(-65) */ L1 = 5.41521234812455953822e-03L, /* high part of ln(2) / 128 */ L2 = 1.31916016021577953656e-17L, /* low part of ln(2) / 128 */ INV_L = 1.84664965233787316146e+02L; /* 128/ln(2) */ /* * The polynomial coefficients are the initial rational approximation * of the minimax polynomial coefficients determined by a Remes algorithm * coded in a much higher working precision. For details, see * * N. Brisebarre, et al, "Computing Machine-Efficient Polynomial * Approximations," ACM Tran. Math. Soft., 32, 236-256 (2006). */ static const long double P1 = 5.0000000000000000000e-1L, P2 = 1.6666666666666429819e-1L, P3 = 4.1666666666666074548e-2L, P4 = 8.3333333333328596382e-3L, P5 = 1.3888888888348716044e-3L, P6 = 1.9841272491100125919e-4L, P7 = 2.4807198720694367822e-5L; #define NUM 128 /* 106-bits of 2^(i/NUM) for i in [0,NUM]. */ static const struct { double hi; double lo; } s[NUM] = { 0x1p+0, 0x0p+0, 0x1.0163da9fb3335p+0, 0x1.b61299ab8cdb7p-54, 0x1.02c9a3e778060p+0, 0x1.dcdef95949ef4p-53, 0x1.04315e86e7f84p+0, 0x1.7ae71f3441b49p-53, 0x1.059b0d3158574p+0, 0x1.d73e2a475b465p-55, 0x1.0706b29ddf6ddp+0, 0x1.8db880753b0f6p-53, 0x1.0874518759bc8p+0, 0x1.186be4bb284ffp-57, 0x1.09e3ecac6f383p+0, 0x1.1487818316136p-54, 0x1.0b5586cf9890fp+0, 0x1.8a62e4adc610bp-54, 0x1.0cc922b7247f7p+0, 0x1.01edc16e24f71p-54, 0x1.0e3ec32d3d1a2p+0, 0x1.03a1727c57b52p-59, 0x1.0fb66affed31ap+0, 0x1.e464123bb1428p-53, 0x1.11301d0125b50p+0, 0x1.49d77e35db263p-53, 0x1.12abdc06c31cbp+0, 0x1.f72575a649ad2p-53, 0x1.1429aaea92ddfp+0, 0x1.66820328764b1p-53, 0x1.15a98c8a58e51p+0, 0x1.2406ab9eeab0ap-55, 0x1.172b83c7d517ap+0, 0x1.b9bef918a1d63p-53, 0x1.18af9388c8de9p+0, 0x1.777ee1734784ap-53, 0x1.1a35beb6fcb75p+0, 0x1.e5b4c7b4968e4p-55, 0x1.1bbe084045cd3p+0, 0x1.3563ce56884fcp-53, 0x1.1d4873168b9aap+0, 0x1.e016e00a2643cp-54, 0x1.1ed5022fcd91cp+0, 0x1.71033fec2243ap-53, 0x1.2063b88628cd6p+0, 0x1.dc775814a8495p-55, 0x1.21f49917ddc96p+0, 0x1.2a97e9494a5eep-55, 0x1.2387a6e756238p+0, 0x1.9b07eb6c70573p-54, 0x1.251ce4fb2a63fp+0, 0x1.ac155bef4f4a4p-55, 0x1.26b4565e27cddp+0, 0x1.2bd339940e9d9p-55, 0x1.284dfe1f56380p+0, 0x1.2d9e2b9e07941p-53, 0x1.29e9df51fdee1p+0, 0x1.612e8afad1255p-55, 0x1.2b87fd0dad98fp+0, 0x1.fbbd48ca71f95p-53, 0x1.2d285a6e4030bp+0, 0x1.0024754db41d5p-54, 0x1.2ecafa93e2f56p+0, 0x1.1ca0f45d52383p-56, 0x1.306fe0a31b715p+0, 0x1.6f46ad23182e4p-55, 0x1.32170fc4cd831p+0, 0x1.a9ce78e18047cp-55, 0x1.33c08b26416ffp+0, 0x1.32721843659a6p-54, 0x1.356c55f929ff0p+0, 0x1.928c468ec6e76p-53, 0x1.371a7373aa9cap+0, 0x1.4e28aa05e8a8fp-53, 0x1.38cae6d05d865p+0, 0x1.0b53961b37da2p-53, 0x1.3a7db34e59ff6p+0, 0x1.d43792533c144p-53, 0x1.3c32dc313a8e4p+0, 0x1.08003e4516b1ep-53, 0x1.3dea64c123422p+0, 0x1.ada0911f09ebcp-55, 0x1.3fa4504ac801bp+0, 0x1.417ee03548306p-53, 0x1.4160a21f72e29p+0, 0x1.f0864b71e7b6cp-53, 0x1.431f5d950a896p+0, 0x1.b8e088728219ap-53, 0x1.44e086061892dp+0, 0x1.89b7a04ef80dp-59, 0x1.46a41ed1d0057p+0, 0x1.c944bd1648a76p-54, 0x1.486a2b5c13cd0p+0, 0x1.3c1a3b69062fp-56, 0x1.4a32af0d7d3dep+0, 0x1.9cb62f3d1be56p-54, 0x1.4bfdad5362a27p+0, 0x1.d4397afec42e2p-56, 0x1.4dcb299fddd0dp+0, 0x1.8ecdbbc6a7833p-54, 0x1.4f9b2769d2ca6p+0, 0x1.5a67b16d3540ep-53, 0x1.516daa2cf6641p+0, 0x1.8225ea5909b04p-53, 0x1.5342b569d4f81p+0, 0x1.be1507893b0d5p-53, 0x1.551a4ca5d920ep+0, 0x1.8a5d8c4048699p-53, 0x1.56f4736b527dap+0, 0x1.9bb2c011d93adp-54, 0x1.58d12d497c7fdp+0, 0x1.295e15b9a1de8p-55, 0x1.5ab07dd485429p+0, 0x1.6324c054647adp-54, 0x1.5c9268a5946b7p+0, 0x1.c4b1b816986a2p-60, 0x1.5e76f15ad2148p+0, 0x1.ba6f93080e65ep-54, 0x1.605e1b976dc08p+0, 0x1.60edeb25490dcp-53, 0x1.6247eb03a5584p+0, 0x1.63e1f40dfa5b5p-53, 0x1.6434634ccc31fp+0, 0x1.8edf0e2989db3p-53, 0x1.6623882552224p+0, 0x1.224fb3c5371e6p-53, 0x1.68155d44ca973p+0, 0x1.038ae44f73e65p-57, 0x1.6a09e667f3bccp+0, 0x1.21165f626cdd5p-53, 0x1.6c012750bdabep+0, 0x1.daed533001e9ep-53, 0x1.6dfb23c651a2ep+0, 0x1.e441c597c3775p-53, 0x1.6ff7df9519483p+0, 0x1.9f0fc369e7c42p-53, 0x1.71f75e8ec5f73p+0, 0x1.ba46e1e5de15ap-53, 0x1.73f9a48a58173p+0, 0x1.7ab9349cd1562p-53, 0x1.75feb564267c8p+0, 0x1.7edd354674916p-53, 0x1.780694fde5d3fp+0, 0x1.866b80a02162dp-54, 0x1.7a11473eb0186p+0, 0x1.afaa2047ed9b4p-53, 0x1.7c1ed0130c132p+0, 0x1.f124cd1164dd6p-54, 0x1.7e2f336cf4e62p+0, 0x1.05d02ba15797ep-56, 0x1.80427543e1a11p+0, 0x1.6c1bccec9346bp-53, 0x1.82589994cce12p+0, 0x1.159f115f56694p-53, 0x1.8471a4623c7acp+0, 0x1.9ca5ed72f8c81p-53, 0x1.868d99b4492ecp+0, 0x1.01c83b21584a3p-53, 0x1.88ac7d98a6699p+0, 0x1.994c2f37cb53ap-54, 0x1.8ace5422aa0dbp+0, 0x1.6e9f156864b27p-54, 0x1.8cf3216b5448bp+0, 0x1.de55439a2c38bp-53, 0x1.8f1ae99157736p+0, 0x1.5cc13a2e3976cp-55, 0x1.9145b0b91ffc5p+0, 0x1.114c368d3ed6ep-53, 0x1.93737b0cdc5e4p+0, 0x1.e8a0387e4a814p-53, 0x1.95a44cbc8520ep+0, 0x1.d36906d2b41f9p-53, 0x1.97d829fde4e4fp+0, 0x1.173d241f23d18p-53, 0x1.9a0f170ca07b9p+0, 0x1.7462137188ce7p-53, 0x1.9c49182a3f090p+0, 0x1.c7c46b071f2bep-56, 0x1.9e86319e32323p+0, 0x1.824ca78e64c6ep-56, 0x1.a0c667b5de564p+0, 0x1.6535b51719567p-53, 0x1.a309bec4a2d33p+0, 0x1.6305c7ddc36abp-54, 0x1.a5503b23e255cp+0, 0x1.1684892395f0fp-53, 0x1.a799e1330b358p+0, 0x1.bcb7ecac563c7p-54, 0x1.a9e6b5579fdbfp+0, 0x1.0fac90ef7fd31p-54, 0x1.ac36bbfd3f379p+0, 0x1.81b72cd4624ccp-53, 0x1.ae89f995ad3adp+0, 0x1.7a1cd345dcc81p-54, 0x1.b0e07298db665p+0, 0x1.2108559bf8deep-53, 0x1.b33a2b84f15fap+0, 0x1.ed7fa1cf7b29p-53, 0x1.b59728de55939p+0, 0x1.1c7102222c90ep-53, 0x1.b7f76f2fb5e46p+0, 0x1.d54f610356a79p-53, 0x1.ba5b030a10649p+0, 0x1.0819678d5eb69p-53, 0x1.bcc1e904bc1d2p+0, 0x1.23dd07a2d9e84p-55, 0x1.bf2c25bd71e08p+0, 0x1.0811ae04a31c7p-53, 0x1.c199bdd85529cp+0, 0x1.11065895048ddp-55, 0x1.c40ab5fffd07ap+0, 0x1.b4537e083c60ap-54, 0x1.c67f12e57d14bp+0, 0x1.2884dff483cadp-54, 0x1.c8f6d9406e7b5p+0, 0x1.1acbc48805c44p-56, 0x1.cb720dcef9069p+0, 0x1.503cbd1e949dbp-56, 0x1.cdf0b555dc3f9p+0, 0x1.889f12b1f58a3p-53, 0x1.d072d4a07897bp+0, 0x1.1a1e45e4342b2p-53, 0x1.d2f87080d89f1p+0, 0x1.15bc247313d44p-53, 0x1.d5818dcfba487p+0, 0x1.2ed02d75b3707p-55, 0x1.d80e316c98397p+0, 0x1.7709f3a09100cp-53, 0x1.da9e603db3285p+0, 0x1.c2300696db532p-54, 0x1.dd321f301b460p+0, 0x1.2da5778f018c3p-54, 0x1.dfc97337b9b5ep+0, 0x1.72d195873da52p-53, 0x1.e264614f5a128p+0, 0x1.424ec3f42f5b5p-53, 0x1.e502ee78b3ff6p+0, 0x1.39e8980a9cc8fp-55, 0x1.e7a51fbc74c83p+0, 0x1.2d522ca0c8de2p-54, 0x1.ea4afa2a490d9p+0, 0x1.0b1ee7431ebb6p-53, 0x1.ecf482d8e67f0p+0, 0x1.1b60625f7293ap-53, 0x1.efa1bee615a27p+0, 0x1.dc7f486a4b6bp-54, 0x1.f252b376bba97p+0, 0x1.3a1a5bf0d8e43p-54, 0x1.f50765b6e4540p+0, 0x1.9d3e12dd8a18bp-54, 0x1.f7bfdad9cbe13p+0, 0x1.1227697fce57bp-53, 0x1.fa7c1819e90d8p+0, 0x1.74853f3a5931ep-55, 0x1.fd3c22b8f71f1p+0, 0x1.2eb74966579e7p-57 }; static const long double tiny = 0x1.p-10000, zero = 0; /* * From Tang's paper: INTRND rounds a floating-point number to the nearest * integer in the manner prescribed by the IEEE standard. It is crucial that * the default round-to-nearest mode, not any other rounding mode, is in * effect here. */ #define INTRND(x) roundl((x)) long double expl(long double x) { union IEEEl2bits z; int j, k, m, n, n1, n2, sgn; long double r1, r2, r, p, q, t; z.e = x; sgn = z.bits.sign; z.bits.sign = 0; /* x is either 0 or a subnormal number. */ if (z.bits.exp == 0) return (1 + x); /* * If x = +Inf, then exp(x) = Inf. * If x = -Inf, then exp(x) = 0. * If x = NaN, then exp(x) = NaN. */ if (z.bits.exp == LDBL_MAX_EXP + BIAS) { mask_nbit_l(z); if (!(z.bits.manh | z.bits.manl)) return (sgn ? zero : 1.L / zero); return ((x - x) / (x - x)); } if (z.e > THRESHOLD_1) return (x > 0 ? (z.e / (z.e - z.e)) : (0 + tiny * tiny)); if (z.e < THRESHOLD_2) return (1 + x); n = INTRND(x * INV_L); n2 = n % NUM; if (n2 < 0) n2 += NUM; n1 = n - n2; if (n >= 512 || n <= -512) r1 = (x - n1 * L1) - n2 * L1; else r1 = x - n * L1; r2 = - n * L2; n1 /= NUM; /* n1 is Tang's m and his j is n2 below. */ r = r1 + r2; q = r * r * (P1 + r * (P2 + r * (P3 + r * (P4 + r * (P5 + r * (P6 + r * P7)))))); p = r1 + (r2 + q); t = (long double)s[n2].lo + s[n2].hi; t = (long double)s[n2].hi + (s[n2].lo + t * p); return (ldexpl(t, n1)); }