This is the mail archive of the
glibc-bugs@sourceware.org
mailing list for the glibc project.
[Bug math/22902] float128 test failures with GCC 8
- From: "joseph at codesourcery dot com" <sourceware-bugzilla at sourceware dot org>
- To: glibc-bugs at sourceware dot org
- Date: Wed, 28 Feb 2018 17:58:43 +0000
- Subject: [Bug math/22902] float128 test failures with GCC 8
- Auto-submitted: auto-generated
- References: <bug-22902-131@http.sourceware.org/bugzilla/>
https://sourceware.org/bugzilla/show_bug.cgi?id=22902
--- Comment #7 from joseph at codesourcery dot com <joseph at codesourcery dot com> ---
The optimized floating-point state handling in fenv_private.h needs to
know which floating-point state - x87 or SSE - is used for each
floating-point type, so that only one state needs updating / testing for
libm code using that state internally.
It seems that the state used by libgcc for float128 cannot be determined
from predefines available when glibc is built, because when glibc is
built for 32-bit with SSE math enabled, libgcc might or might not have
been built that way, and the way libgcc was built affects which state is
used for float128. Indeed, it would also be possible for libgcc to be
built with SSE math but glibc without.
Thus, this indicates that the optimized functions cannot be used for
float128 for 32-bit x86. Instead, libc_feholdexcept_setroundf128 and
libc_feupdateenv_testf128 should, I suppose, be defined to
default_libc_feholdexcept_setround and default_libc_feupdateenv_test for
32-bit x86, if that works to fix this issue.
--
You are receiving this mail because:
You are on the CC list for the bug.