This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: Obsolete matherr, _LIB_VERSION, libieee.a
- From: Szabolcs Nagy <szabolcs dot nagy at arm dot com>
- To: sellcey at cavium dot com, Joseph Myers <joseph at codesourcery dot com>, libc-alpha at sourceware dot org
- Cc: nd at arm dot com
- Date: Thu, 24 Aug 2017 10:19:53 +0100
- Subject: Re: Obsolete matherr, _LIB_VERSION, libieee.a
- Authentication-results: sourceware.org; auth=none
- Authentication-results: spf=none (sender IP is ) smtp.mailfrom=Szabolcs dot Nagy at arm dot com;
- Nodisclaimer: True
- References: <alpine.DEB.2.20.1708172128090.26254@digraph.polyomino.org.uk> <1503519708.28672.77.camel@cavium.com>
- Spamdiagnosticmetadata: NSPM
- Spamdiagnosticoutput: 1:99
On 23/08/17 21:21, Steve Ellcey wrote:
> On Thu, 2017-08-17 at 21:29 +0000, Joseph Myers wrote:
>> This patch obsoletes support for SVID libm error handling (the system
>> where a user-defined function matherr is called on a libm function
>> error; only enabled if you also set _LIB_VERSION = _SVID_ or
>> _LIB_VERSION = _XOPEN_) and the use of the _LIB_VERSION global
>> variable to control libm error handling.
>
> Joseph,
>
> I am having a problem with this patch when combined with my aarch64
> ILP32 patches. The build is fine but during ILP32 testing I get:
>
> /toolchain-ilp32/obj/glibc32/math/test-matherr.o: In function `do_test':
> /toolchain-ilp32/src/glibc/math/test-matherr.c:43: undefined reference to `_LIB_VERSION@GLIBC_2.27'
> /toolchain-ilp32/src/glibc/math/test-matherr.c:43: undefined reference to `_LIB_VERSION@GLIBC_2.27'
>
> I know the ILP32 changes aren't in the official tree yet but I was
> wondering if there is some way to fix this. I assume the problem is
> because there was never an aarch64 ILP32 that supported matherr.
>
> I actually have a similar problem with tst-mallocstate and the
> malloc_get_state symbol. Currently I just remove that test from the makefile
> in order to skip it and I guess I can do the same for this test but we
> will need a solution to this problem at some point and I was wondering
> if you had any ideas.
>
> Earlier discussion about tst-mallocstate:
> https://sourceware.org/ml/libc-alpha/2016-12/msg00527.html
>
generating an abi-versions.h sounds reasonable.
a lazier solution would be to add an 'unsupported-tests'
make var that targets could set and get filtered out.