This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH][BZ #16823] Fix log1pl returning wrong infinity sign
- From: David Miller <davem at davemloft dot net>
- To: stli at linux dot vnet dot ibm dot com
- Cc: libc-alpha at sourceware dot org
- Date: Tue, 22 Apr 2014 20:42:03 -0400 (EDT)
- Subject: Re: [PATCH][BZ #16823] Fix log1pl returning wrong infinity sign
- Authentication-results: sourceware.org; auth=none
- References: <li3hm5$hus$1 at ger dot gmane dot org> <53562754 dot 3040706 at linux dot vnet dot ibm dot com>
From: Stefan Liebler <stli@linux.vnet.ibm.com>
Date: Tue, 22 Apr 2014 10:24:52 +0200
> log1pl (-1) now returns with -inf instead of +inf in FE_DOWNWARD
> rounding mode. The source code producing the -inf for the double/float
> variants is the same, but on S390 the compiler generates a division by
> constant +zero and ignores the (x-x).
> Thus i also changed it to the constant zero in the double/float
> functions.
> I´ve tested it on S390.
> Please test it on other platforms because there are also reports for
> the same failure on aarch64 (see bug comment 1)
> and sparc
> (https://www.sourceware.org/ml/libc-alpha/2014-04/msg00410.html).
>
> Give feedback in order to commit the patch.
I can confirm that this fixes the testsuite failures on sparc.