diff options
| author | Joseph Myers <joseph@codesourcery.com> | 2015-05-15 17:47:29 +0000 |
|---|---|---|
| committer | Joseph Myers <joseph@codesourcery.com> | 2015-05-15 17:47:29 +0000 |
| commit | 5a608ccc2d676aa124f6b55a7b20d071e408ca31 (patch) | |
| tree | 41c40103e1b049d25bc519ca0688f48d82d61656 /sysdeps | |
| parent | ff069f024ae8cf15d53429e034d67ddcece0f67a (diff) | |
| download | glibc-5a608ccc2d676aa124f6b55a7b20d071e408ca31.tar glibc-5a608ccc2d676aa124f6b55a7b20d071e408ca31.tar.gz glibc-5a608ccc2d676aa124f6b55a7b20d071e408ca31.tar.bz2 glibc-5a608ccc2d676aa124f6b55a7b20d071e408ca31.zip | |
Fix tanf spurious underflows (bug 18221).
The flt-32 implementation of tanf produces spurious underflow
exceptions for some small arguments, through computing values on the
order of x^5. This patch fixes this by adjusting the threshold for
returning x (or, as applicable, +/- 1/x) to 2**-13 (the next term in
the power series being x^3/3).
Tested for x86_64 and x86.
[BZ #18221]
* sysdeps/ieee754/flt-32/k_tanf.c (__kernel_tanf): Use 2**-13 not
2**-28 as threshold for returning x or +/- 1/x.
* math/auto-libm-test-in: Add more tests of tan.
* math/auto-libm-test-out: Regenerated.
Diffstat (limited to 'sysdeps')
| -rw-r--r-- | sysdeps/ieee754/flt-32/k_tanf.c | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/sysdeps/ieee754/flt-32/k_tanf.c b/sysdeps/ieee754/flt-32/k_tanf.c index d918826338..a67f36e283 100644 --- a/sysdeps/ieee754/flt-32/k_tanf.c +++ b/sysdeps/ieee754/flt-32/k_tanf.c @@ -45,7 +45,7 @@ float __kernel_tanf(float x, float y, int iy) int32_t ix,hx; GET_FLOAT_WORD(hx,x); ix = hx&0x7fffffff; /* high word of |x| */ - if(ix<0x31800000) /* x < 2**-28 */ + if(ix<0x39000000) /* x < 2**-13 */ {if((int)x==0) { /* generate inexact */ if((ix|(iy+1))==0) return one/fabsf(x); else return (iy==1)? x: -one/x; |