https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71806
Bug ID: 71806
Summary: PowerPC -mcpu=power9 enables __float128 without an
explicit -mfloat128
Product: gcc
Version: 6.1.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: target
Assignee: unassigned at gcc dot gnu.org
Reporter: meissner at gcc dot gnu.org
Target Milestone: ---
The macro ISA_3_0_MASKS_SERVER includes OPTION_MASK_FLOAT128_HW, which enables
the -mfloat128 option. The -mfloat128 option being on by default means that
some of the libraries (libstdc++, libquadmath, etc.) will add code for IEEE
128-bit floating point.
Until all of the libraries are updated to support IEEE 128-bit, the
-mcpu=power9 support should not enable IEEE 128-bit floating point support
(either hardware or software).
In addition, if you configure the compiler with the --with-cpu=power9 option,
it should not enable IEEE 128-bit floating point support by default.
If the user uses -mcpu=power9 -mfloat128, it should enable the hardware support
that is available in ISA 3.0.
