Fix narrowing function tests build for powerpc64le.
commita85b70dbd2efb84123dcfb7a2131b2de81b59b6a
authorJoseph Myers <joseph@codesourcery.com>
Sat, 10 Feb 2018 01:52:33 +0000 (10 01:52 +0000)
committerJoseph Myers <joseph@codesourcery.com>
Sat, 10 Feb 2018 01:52:33 +0000 (10 01:52 +0000)
tree632f77b7e0edd5c5fa01234de2d6bcd95e597866
parent6a1ff640dcec04905d8518983ad6252d38b7a733
Fix narrowing function tests build for powerpc64le.

Testing narrowing functions with build-many-glibcs.py showed up a
further testsuite fix needed to enable building such functions for
powerpc64le: tests test-<narrower-type>-float128-<function> (and
likewise for float64x) needed the same special handling for
powerpc64le as test-float128-* and test-float64x-*.  This patch adds
that special handling.

Tested with build-many-glibcs.py for powerpc64le in conjunction with
the main patch adding narrowing add functions.

* sysdeps/powerpc/powerpc64le/Makefile [$(subdir) = math]
(f128-pairs): New variable.
[$(subdir) = math] ($(foreach suf,$(all-object-suffixes),$(foreach
pair,$(f128-pairs),$(objpfx)test-$(pair)%$(suf)))): Add -mfloat128
to CFLAGS.
[$(subdir) = math] ($(foreach pair,$(f128-pairs),test-$(pair)%)):
Also make tests add $(f128-loader-link) to gnulib-tests.
ChangeLog
sysdeps/powerpc/powerpc64le/Makefile