Bug 18248 - missing underflow for subnormal inputs in sin/sinh/tan/tanh/atanh/expm1/log1p/exp2
Summary: missing underflow for subnormal inputs in sin/sinh/tan/tanh/atanh/expm1/log1...
Status: RESOLVED DUPLICATE of bug 16339
Alias: None
Product: glibc
Classification: Unclassified
Component: math (show other bugs)
Version: unspecified
: P2 normal
Target Milestone: ---
Assignee: Not yet assigned to anyone
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2015-04-10 13:11 UTC by Szabolcs Nagy
Modified: 2015-04-20 07:39 UTC (History)
0 users

See Also:
Host:
Target:
Build:
Last reconfirmed:
fweimer: security-

Attachments
underflow tests (383 bytes, text/x-csrc)
2015-04-10 13:11 UTC, Szabolcs Nagy 		Details
View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Szabolcs Nagy 2015-04-10 13:11:18 UTC 
Created attachment 8238 [details]
underflow tests

some functions don't raise the underflow flag for subnormal inputs (even though the result is inexact subnormal)

attach code on x86_64 gives:

sin(0x1p-1030) = 0x8p-1033 = 8.69169e-311 did not underflow
sin(-0x1p-1030) = -0x8p-1033 = -8.69169e-311 did not underflow
sinl(0x1p-16400L) = 0x0.0002p-16385 = 1.28254e-4937 did not underflow
sinl(-0x1p-16400L) = -0x0.0002p-16385 = -1.28254e-4937 did not underflow
sinh(0x1p-1030) = 0x8p-1033 = 8.69169e-311 did not underflow
sinh(-0x1p-1030) = -0x8p-1033 = -8.69169e-311 did not underflow
sinhf(0x1p-130f) = 0x8p-133 = 7.34684e-40 did not underflow
sinhf(-0x1p-130f) = -0x8p-133 = -7.34684e-40 did not underflow
sinhl(0x1p-16400L) = 0x0.0002p-16385 = 1.28254e-4937 did not underflow
sinhl(-0x1p-16400L) = -0x0.0002p-16385 = -1.28254e-4937 did not underflow
tan(0x1p-1030) = 0x8p-1033 = 8.69169e-311 did not underflow
tan(-0x1p-1030) = -0x8p-1033 = -8.69169e-311 did not underflow
tanf(0x1p-130f) = 0x8p-133 = 7.34684e-40 did not underflow
tanf(-0x1p-130f) = -0x8p-133 = -7.34684e-40 did not underflow
tanl(0x1p-16400L) = 0x0.0002p-16385 = 1.28254e-4937 did not underflow
tanl(-0x1p-16400L) = -0x0.0002p-16385 = -1.28254e-4937 did not underflow
tanh(0x1p-1030) = 0x8p-1033 = 8.69169e-311 did not underflow
tanh(-0x1p-1030) = -0x8p-1033 = -8.69169e-311 did not underflow
tanhf(0x1p-130f) = 0x8p-133 = 7.34684e-40 did not underflow
tanhf(-0x1p-130f) = -0x8p-133 = -7.34684e-40 did not underflow
tanhl(0x1p-16400L) = 0x0.0002p-16385 = 1.28254e-4937 did not underflow
tanhl(-0x1p-16400L) = -0x0.0002p-16385 = -1.28254e-4937 did not underflow
atanh(0x1p-1030) = 0x8p-1033 = 8.69169e-311 did not underflow
atanh(-0x1p-1030) = -0x8p-1033 = -8.69169e-311 did not underflow
atanhf(0x1p-130f) = 0x8p-133 = 7.34684e-40 did not underflow
atanhf(-0x1p-130f) = -0x8p-133 = -7.34684e-40 did not underflow
atanhl(0x1p-16400L) = 0x0.0002p-16385 = 1.28254e-4937 did not underflow
atanhl(-0x1p-16400L) = -0x0.0002p-16385 = -1.28254e-4937 did not underflow
expm1(0x1p-1030) = 0x8p-1033 = 8.69169e-311 did not underflow
expm1(-0x1p-1030) = -0x8p-1033 = -8.69169e-311 did not underflow
expm1f(0x1p-130f) = 0x8p-133 = 7.34684e-40 did not underflow
expm1f(-0x1p-130f) = -0x8p-133 = -7.34684e-40 did not underflow
log1p(0x1p-1030) = 0x8p-1033 = 8.69169e-311 did not underflow
log1p(-0x1p-1030) = -0x8p-1033 = -8.69169e-311 did not underflow
log1pf(0x1p-130f) = 0x8p-133 = 7.34684e-40 did not underflow
log1pf(-0x1p-130f) = -0x8p-133 = -7.34684e-40 did not underflow
exp2l(-16382.5) = 0x5.a827999fcef3242p-16385 = 2.37737e-4932 did not underflow


on aarch64 it gives:

sin(0x1p-1030) = 0x1p-1030 = 8.69169e-311 did not underflow
sin(-0x1p-1030) = -0x1p-1030 = -8.69169e-311 did not underflow
sinf(0x1p-130f) = 0x1p-130 = 7.34684e-40 did not underflow
sinf(-0x1p-130f) = -0x1p-130 = -7.34684e-40 did not underflow
sinl(0x1p-16400L) = 0x0.00004p-16382 = 1.28254e-4937 did not underflow
sinl(-0x1p-16400L) = -0x0.00004p-16382 = -1.28254e-4937 did not underflow
sinh(0x1p-1030) = 0x1p-1030 = 8.69169e-311 did not underflow
sinh(-0x1p-1030) = -0x1p-1030 = -8.69169e-311 did not underflow
sinhf(0x1p-130f) = 0x1p-130 = 7.34684e-40 did not underflow
sinhf(-0x1p-130f) = -0x1p-130 = -7.34684e-40 did not underflow
sinhl(0x1p-16400L) = 0x0.00004p-16382 = 1.28254e-4937 did not underflow
sinhl(-0x1p-16400L) = -0x0.00004p-16382 = -1.28254e-4937 did not underflow
tan(0x1p-1030) = 0x1p-1030 = 8.69169e-311 did not underflow
tan(-0x1p-1030) = -0x1p-1030 = -8.69169e-311 did not underflow
tanf(0x1p-130f) = 0x1p-130 = 7.34684e-40 did not underflow
tanf(-0x1p-130f) = -0x1p-130 = -7.34684e-40 did not underflow
tanl(0x1p-16400L) = 0x0.00004p-16382 = 1.28254e-4937 did not underflow
tanl(-0x1p-16400L) = -0x0.00004p-16382 = -1.28254e-4937 did not underflow
tanh(0x1p-1030) = 0x1p-1030 = 8.69169e-311 did not underflow
tanh(-0x1p-1030) = -0x1p-1030 = -8.69169e-311 did not underflow
tanhf(0x1p-130f) = 0x1p-130 = 7.34684e-40 did not underflow
tanhf(-0x1p-130f) = -0x1p-130 = -7.34684e-40 did not underflow
tanhl(0x1p-16400L) = 0x0.00004p-16382 = 1.28254e-4937 did not underflow
tanhl(-0x1p-16400L) = -0x0.00004p-16382 = -1.28254e-4937 did not underflow
atanh(0x1p-1030) = 0x1p-1030 = 8.69169e-311 did not underflow
atanh(-0x1p-1030) = -0x1p-1030 = -8.69169e-311 did not underflow
atanhf(0x1p-130f) = 0x1p-130 = 7.34684e-40 did not underflow
atanhf(-0x1p-130f) = -0x1p-130 = -7.34684e-40 did not underflow
atanhl(0x1p-16400L) = 0x0.00004p-16382 = 1.28254e-4937 did not underflow
atanhl(-0x1p-16400L) = -0x0.00004p-16382 = -1.28254e-4937 did not underflow
expm1(0x1p-1030) = 0x1p-1030 = 8.69169e-311 did not underflow
expm1(-0x1p-1030) = -0x1p-1030 = -8.69169e-311 did not underflow
expm1f(0x1p-130f) = 0x1p-130 = 7.34684e-40 did not underflow
expm1f(-0x1p-130f) = -0x1p-130 = -7.34684e-40 did not underflow
log1p(0x1p-1030) = 0x1p-1030 = 8.69169e-311 did not underflow
log1p(-0x1p-1030) = -0x1p-1030 = -8.69169e-311 did not underflow
log1pf(0x1p-130f) = 0x1p-130 = 7.34684e-40 did not underflow
log1pf(-0x1p-130f) = -0x1p-130 = -7.34684e-40 did not underflow
log1pl(0x1p-16400L) = 0x0.00004p-16382 = 1.28254e-4937 did not underflow
log1pl(-0x1p-16400L) = -0x0.00004p-16382 = -1.28254e-4937 did not underflow
Comment 1 Joseph Myers 2015-04-10 16:43:48 UTC 
Omnibus reports for lots of unrelated functions are unhelpful.  This is a duplicate of 16538 (sin), 16519 (sinh), 16517 (tan), 16520 (tanh), 16352 (atanh), 16353 (expm1), 16339 (log1p), 16521 (exp2).

*** This bug has been marked as a duplicate of bug 16339 ***