This is the mail archive of the libc-alpha@sourceware.org mailing list for the glibc project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Avoid ordered comparisons of NaNs in ldbl-128ibm acosl and asinl


On 10/10/2013 07:23 PM, Joseph S. Myers wrote:
> ldbl-128ibm acosl and asinl receive arguments that may be NaN (the
> wrappers only check for arguments with magnitude > 1.0), and do
> ordered comparisons on them.  The only reason this fails to raise
> spurious "invalid" exceptions is, for hard-float, a GCC bug, which
> I've filed as <http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58684>.
> For soft-float, it's a soft-fp bug, bug 14910, which I've proposed a
> patch <https://sourceware.org/ml/libc-alpha/2013-10/msg00365.html> to
> fix.
> 
> This patch adds a check for NaNs at the start of those functions to
> avoid the problem (both for soft-float now, and for hard-float if the
> GCC bug is fixed in future).
> 
> Tested for powerpc-nofpu.
> 
> 2013-10-10  Joseph Myers  <joseph@codsourcery.com>
> 
> 	* sysdeps/ieee754/ldbl-128ibm/e_acosl.c (__ieee754_acosl): Check
> 	for NaNs before doing comparisons on argument.
> 	* sysdeps/ieee754/ldbl-128ibm/e_asinl.c (__ieee754_asinl):
> 	Likewise.

Ok,

Andreas
-- 
 Andreas Jaeger aj@{suse.com,opensuse.org} Twitter/Identica: jaegerandi
  SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany
   GF: Jeff Hawn,Jennifer Guild,Felix Imendörffer,HRB16746 (AG Nürnberg)
    GPG fingerprint = 93A3 365E CE47 B889 DF7F  FED1 389A 563C C272 A126


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]