glibc/sysdeps/ieee754/ldbl-128
Joseph Myers 4629c866ad Fix atan / atan2 missing underflows (bug 15319).
This patch fixes bug 15319, missing underflows from atan / atan2 when
the result of atan is very close to its small argument (or that of
atan2 is very close to the ratio of its arguments, which may be an
exact division).

The usual approach of doing an underflowing computation if the
computed result is subnormal is followed.  For 32-bit x86, there are
extra complications: the inline __ieee754_atan2 in bits/mathinline.h
needs to be disabled for float and double because other libm functions
using it generally rely on getting proper underflow exceptions from
it, while the out-of-line functions have to remove excess range and
precision from the underflowing result so as to return an exact 0 in
the case where errno should be set for underflow to 0.  (The failures
I saw without that are similar to those Carlos reported for other
functions, where I haven't seen a response to
<https://sourceware.org/ml/libc-alpha/2015-01/msg00485.html>
confirming if my diagnosis is correct.  Arguably all libm functions
with float and double returns should remove excess range and
precision, but that's a separate matter.)

The x86_64 long double case reported in a comment in bug 15319 is not
a bug (it's an argument of LDBL_MIN, and x86_64 is an after-rounding
architecture so the correct IEEE result is not to raise underflow in
the given rounding mode, in addition to treating the result as an
exact LDBL_MIN being within the newly clarified documentation of
accuracy goals).  I'm presuming that the fpatan instruction can be
trusted to raise appropriate exceptions when the (long double) result
underflows (after rounding) and so no changes are needed for x86 /
x86_64 long double functions here; empirically this is the case for
the cases covered in the testsuite, on my system.

Tested for x86_64, x86, powerpc and mips64.  Only 32-bit x86 needs
ulps updates (for the changes to inlines meaning some functions no
longer get excess precision from their __ieee754_atan2* calls).

	[BZ #15319]
	* sysdeps/i386/fpu/e_atan2.S (dbl_min): New object.
	(MO): New macro.
	(__ieee754_atan2): For results with small absolute value, force
	underflow exception and remove excess range and precision from
	return value.
	* sysdeps/i386/fpu/e_atan2f.S (flt_min): New object.
	(MO): New macro.
	(__ieee754_atan2f): For results with small absolute value, force
	underflow exception and remove excess range and precision from
	return value.
	* sysdeps/i386/fpu/s_atan.S (dbl_min): New object.
	(MO): New macro.
	(__atan): For results with small absolute value, force underflow
	exception and remove excess range and precision from return value.
	* sysdeps/i386/fpu/s_atanf.S (flt_min): New object.
	(MO): New macro.
	(__atanf): For results with small absolute value, force underflow
	exception and remove excess range and precision from return value.
	* sysdeps/ieee754/dbl-64/e_atan2.c: Include <float.h> and
	<math.h>.
	(__ieee754_atan2): Force underflow exception for results with
	small absolute value.
	* sysdeps/ieee754/dbl-64/s_atan.c: Include <float.h> and
	<math_private.h>.
	(atan): Force underflow exception for results with small absolute
	value.
	* sysdeps/ieee754/flt-32/s_atanf.c: Include <float.h>.
	(__atanf): Force underflow exception for results with small
	absolute value.
	* sysdeps/ieee754/ldbl-128/s_atanl.c: Include <float.h> and
	<math.h>.
	(__atanl): Force underflow exception for results with small
	absolute value.
	* sysdeps/ieee754/ldbl-128ibm/s_atanl.c: Include <float.h>.
	(__atanl): Force underflow exception for results with small
	absolute value.
	* sysdeps/x86/fpu/bits/mathinline.h
	[!__SSE2_MATH__ && !__x86_64__ && __LIBC_INTERNAL_MATH_INLINES]
	(__ieee754_atan2): Only define inline for long double.
	* sysdeps/x86_64/fpu/multiarch/e_atan2.c
	[HAVE_FMA4_SUPPORT || HAVE_AVX_SUPPORT]: Include <math.h>.
	* math/auto-libm-test-in: Do not mark underflow exceptions as
	possibly missing for bug 15319.  Add more tests of atan2.
	* math/auto-libm-test-out: Regenerated.
	* math/libm-test.inc (casin_test_data): Do not mark underflow
	exceptions as possibly missing for bug 15319.
	(casinh_test_data): Likewise.
	* sysdeps/i386/fpu/libm-test-ulps: Update.
2015-02-18 21:10:49 +00:00
..
e_acoshl.c Use <> for math.h and math_private.h everywhere. 2012-03-09 16:09:10 -08:00
e_acosl.c Use <> for math.h and math_private.h everywhere. 2012-03-09 16:09:10 -08:00
e_asinl.c Use <> for math.h and math_private.h everywhere. 2012-03-09 16:09:10 -08:00
e_atan2l.c Use <> for math.h and math_private.h everywhere. 2012-03-09 16:09:10 -08:00
e_atanhl.c Use <> for math.h and math_private.h everywhere. 2012-03-09 16:09:10 -08:00
e_coshl.c Fix cosh spurious underflows from expm1 (bug 16354), inaccurate results near 0 (bug 17061). 2014-06-23 20:20:10 +00:00
e_exp10l.c Update copyright dates with scripts/update-copyrights. 2015-01-02 16:29:47 +00:00
e_expl.c Update copyright dates with scripts/update-copyrights. 2015-01-02 16:29:47 +00:00
e_fmodl.c Use <> for math.h and math_private.h everywhere. 2012-03-09 16:09:10 -08:00
e_gammal_r.c Update copyright dates with scripts/update-copyrights. 2015-01-02 16:29:47 +00:00
e_hypotl.c Fix hypot handling of subnormals (bug 16316, bug 16330). 2013-12-17 13:42:13 +00:00
e_ilogbl.c Finish ilogb changes 2012-04-18 14:31:43 +02:00
e_j0l.c Fix Bessel function spurious overflows for ldbl-128 / ldbl-128ibm (bug 15285). 2013-03-21 13:57:21 +00:00
e_j1l.c Set errno for y1 overflow (bug 17050). 2014-06-23 20:17:13 +00:00
e_jnl.c Fix yn overflow handling in non-default rounding modes (bug 16561, bug 16562). 2014-06-27 14:52:13 +00:00
e_lgammal_r.c powerpc: Fix lgammal_r overflow warnings 2014-12-11 07:17:11 -05:00
e_log2l.c Fix log2 (1) in round-downward mode (bug 17042). 2014-06-10 12:07:15 +00:00
e_log10l.c Fix log10 (1) in round-downward mode (bug 16977). 2014-05-23 12:07:50 +00:00
e_logl.c Fix log (1) in round-downward mode (bug 16731). 2014-03-21 18:13:58 +00:00
e_powl.c Fix ldbl-128 powl sign of result in overflow / underflow cases (bug 17097). 2014-06-29 11:49:08 +00:00
e_rem_pio2l.c Update copyright dates with scripts/update-copyrights. 2015-01-02 16:29:47 +00:00
e_remainderl.c Use <> for math.h and math_private.h everywhere. 2012-03-09 16:09:10 -08:00
e_sinhl.c Use <> for math.h and math_private.h everywhere. 2012-03-09 16:09:10 -08:00
gamma_productl.c Update copyright dates with scripts/update-copyrights. 2015-01-02 16:29:47 +00:00
ieee754.h Update copyright dates with scripts/update-copyrights. 2015-01-02 16:29:47 +00:00
k_cosl.c Update copyright dates with scripts/update-copyrights. 2015-01-02 16:29:47 +00:00
k_sincosl.c Update copyright dates with scripts/update-copyrights. 2015-01-02 16:29:47 +00:00
k_sinl.c Update copyright dates with scripts/update-copyrights. 2015-01-02 16:29:47 +00:00
k_tanl.c Remove trailing whitespace. 2013-06-05 20:44:03 +00:00
ldbl2mpn.c Update copyright dates with scripts/update-copyrights. 2015-01-02 16:29:47 +00:00
Makefile * sysdeps/ieee754/ldbl-128/Makefile: New file. 2007-03-14 17:36:18 +00:00
math_ldbl.h Update. 1999-07-15 18:26:25 +00:00
mpn2ldbl.c Update copyright dates with scripts/update-copyrights. 2015-01-02 16:29:47 +00:00
printf_fphex.c Update copyright dates with scripts/update-copyrights. 2015-01-02 16:29:47 +00:00
s_asinhl.c Use <> for math.h and math_private.h everywhere. 2012-03-09 16:09:10 -08:00
s_atanl.c Fix atan / atan2 missing underflows (bug 15319). 2015-02-18 21:10:49 +00:00
s_cbrtl.c Use <> for math.h and math_private.h everywhere. 2012-03-09 16:09:10 -08:00
s_ceill.c Use <> for math.h and math_private.h everywhere. 2012-03-09 16:09:10 -08:00
s_copysignl.c Use <> for math.h and math_private.h everywhere. 2012-03-09 16:09:10 -08:00
s_cosl.c Use <> for math.h and math_private.h everywhere. 2012-03-09 16:09:10 -08:00
s_erfl.c Fix ldbl-128 erfl spurious underflows (bug 16287). 2014-06-24 20:56:56 +00:00
s_expm1l.c Fix ldbl-128 expm1l spurious underflow (bug 16539). 2014-06-30 17:38:16 +00:00
s_fabsl.c Use <> for math.h and math_private.h everywhere. 2012-03-09 16:09:10 -08:00
s_finitel.c Use <> for math.h and math_private.h everywhere. 2012-03-09 16:09:10 -08:00
s_floorl.c Use <> for math.h and math_private.h everywhere. 2012-03-09 16:09:10 -08:00
s_fma.c Update copyright dates with scripts/update-copyrights. 2015-01-02 16:29:47 +00:00
s_fmal.c Update copyright dates with scripts/update-copyrights. 2015-01-02 16:29:47 +00:00
s_fpclassifyl.c Update copyright dates with scripts/update-copyrights. 2015-01-02 16:29:47 +00:00
s_frexpl.c Use <> for math.h and math_private.h everywhere. 2012-03-09 16:09:10 -08:00
s_isinf_nsl.c Use <> for math.h and math_private.h everywhere. 2012-03-09 16:09:10 -08:00
s_isinfl.c Use <> for math.h and math_private.h everywhere. 2012-03-09 16:09:10 -08:00
s_isnanl.c Use <> for math.h and math_private.h everywhere. 2012-03-09 16:09:10 -08:00
s_issignalingl.c Update copyright dates with scripts/update-copyrights. 2015-01-02 16:29:47 +00:00
s_llrintl.c Update copyright dates with scripts/update-copyrights. 2015-01-02 16:29:47 +00:00
s_llroundl.c Update copyright dates with scripts/update-copyrights. 2015-01-02 16:29:47 +00:00
s_log1pl.c Fix log1pl (LDBL_MAX) in FE_UPWARD mode (bug 16564). 2014-05-14 12:38:56 +00:00
s_logbl.c Optimize handling of denormals in logb/logbf/logbl 2012-05-26 13:53:22 +02:00
s_lrintl.c Update copyright dates with scripts/update-copyrights. 2015-01-02 16:29:47 +00:00
s_lroundl.c Update copyright dates with scripts/update-copyrights. 2015-01-02 16:29:47 +00:00
s_modfl.c Use <> for math.h and math_private.h everywhere. 2012-03-09 16:09:10 -08:00
s_nearbyintl.c Fix nearbyint scheduling of arithmetic past fesetenv (bug 15490). 2013-05-19 18:40:25 +00:00
s_nextafterl.c Fix nextafter overflow in non-default rounding modes (bug 16677). 2014-03-11 22:24:00 +00:00
s_nexttoward.c Fix leading whitespaces. 2013-06-06 20:36:07 +02:00
s_nexttowardf.c Remove trailing whitespace. 2013-06-05 20:44:03 +00:00
s_remquol.c Fix sign of remquo zero remainder in round-downward mode (bug 17987). 2015-02-17 00:41:50 +00:00
s_rintl.c Fix set-but-not-used warnings in ldbl-128 nearbyintl, rintl. 2012-11-20 14:26:07 +00:00
s_roundl.c Update copyright dates with scripts/update-copyrights. 2015-01-02 16:29:47 +00:00
s_scalblnl.c [BZ #6803] Set errno for scalbln, scalbn 2014-06-20 07:48:20 +05:30
s_scalbnl.c Remove trailing whitespace. 2013-06-05 20:44:03 +00:00
s_signbitl.c Update copyright dates with scripts/update-copyrights. 2015-01-02 16:29:47 +00:00
s_sincosl.c Fix sincos errno setting (bug 15467). 2015-02-11 23:17:25 +00:00
s_sinl.c Use <> for math.h and math_private.h everywhere. 2012-03-09 16:09:10 -08:00
s_tanhl.c Use <> for math.h and math_private.h everywhere. 2012-03-09 16:09:10 -08:00
s_tanl.c Remove trailing whitespace. 2013-06-05 20:44:03 +00:00
s_truncl.c Update copyright dates with scripts/update-copyrights. 2015-01-02 16:29:47 +00:00
strtold_l.c Update copyright dates with scripts/update-copyrights. 2015-01-02 16:29:47 +00:00
t_expl.h Update copyright dates with scripts/update-copyrights. 2015-01-02 16:29:47 +00:00
t_sincosl.c Update copyright dates with scripts/update-copyrights. 2015-01-02 16:29:47 +00:00
w_expl.c [BZ #16427] Fix ldbl-128 exp overflows. 2014-01-15 09:50:31 +01:00
x2y2m1l.c Update copyright dates with scripts/update-copyrights. 2015-01-02 16:29:47 +00:00